Guaranteed reliable broadcast network.

[Detailed Description of the Invention]

It is related with the broadcast network where reliability was guaranteed. It can classify into a concept. - 1 to 1 - It is pair 1 in large numbers. - One-pair large number It is alike and explains. Two sets of two sets of personal pair Eta which communicate mutually as an example using a telephone line, and the digital data terminals which communicate via a satellite link can be mentioned. When a data message is transmitted toward outside from a transmitting side data processing station, If interfere electrically, connection loosens, hardware breaks down, the processing speed of a receiving side data processing station is slow or a receiving side data processing station is in use, A data message is lost or it may distort (for example, 1 changes to O). Then, the communications protocol is used in order to guarantee exact reception of data. Art of returning the message which shows the receiving side data processing station having received the message or the message of - group correctly to a communications protocol to a transmitting side data processing station, What a transmitting side data processing station broadcasts a message or the message of - group again by notice, and says C That a receiving side data-processing stage Reason station transmits a message, and keeps at a speed earlier than the processing speed of a receiving side data processing station is prevented certainly immediately. A transmitting side station is [ Which message was actually received by the receiving side station correctly? ] that of "knowing", this message advice-of-receipt method functions very well, and it guarantees reliable data communications in communication of 1 to 1. They are the data communications of pair 1 in large numbers. The processing in the case of transmitting a message from many data processing stations to one set of a certain data processing station does not have double Problem more (this is because many data processing stations are functioning independently mutually, for example). There are some which have the capability of Lucic cast (for example, a token ring, Ethernet, radio, microwave, a satellite network, etc.). However, in a publicly known system, multicasting or the broadcast message wipes a possibility of being accidentally received incorrectly by one or more network nodes, and does not go out. (For example, it can interfere electrically, connection can loosen, hardware, the processing speed of a receiving side data processing station can be slow, or a receiving side data processing station is in use, and a message can distort). Since the network communication hardware currently installed independently has sometimes made the error for a certain reason also in the network designed most carefully, it is not a complete guarantee whether data is received correctly. Mr. Metcalfue and others [ for example, ] (Metcal f'e et al) of Xerox Park (XEROX PARC), Important early paper "Ethernet: Distributed packet change [ for local computer networks ] (Ethernet: Dls tributed Packet SwitchingFor Local C omputer Networks) J, A communications 19 of CM -- the inside of No. 7 (July, 1976) -- other packets (19 Communicationsof th e ACM No and 7 (July 1976)) It is warning that an Ethernet packet will be lost for the non-operativeness of Interference, a noise, and a receiving side station, and other reasons. It is said that the protocol used for communication according [ the author ] to Ethernet must assume that it is what is received correctly (however, it is not certain) by the probability that a packet is high in the addressee. The author has secured communicative reliability by obliging it to follow the required step which exists in the case of transmission and addressee specification processing. In Communications Act of existing one-pair a large number, if the guarantee of data reception is not offered, either, the art which notifies a sending node whether reception by each receiving node is exact or inaccurate is not provided, either. Therefore, Communications Act of existing one-pair a large number is known as J services of [datagram (datagram). It is not efficient to guarantee exact reception of data to communication of one-pair a large number using the art of message advice of receipt useful to communication of 1 to 1. For example, it is supposing one set of a certain data processing station must transmit the same data to 1,000 sets of other data processing stations and it must wait for 1,000 advice-of-receipt messages, When it is at the time when some stations in 1.000 sets of data processing stations are busy by chance or is not operating, it must wait very long time. Nevertheless, when the reliability of the data communications by broadcast or multicasting needs to be secured, most uses Communications Act of 1 to 1 now. For example, when the same contents of communication must be transmitted to 1,000 sets of other data processing stations, in order to secure reliable data communications between each station, the data communications of 1 to 1 are performed no less than 1.000 times. This method secures reliability at the sacrifice of efficiency. They are the data communications of opposite a large number in large numbers. It is not practical that a large number return to communication of 1 to 1 in communication of opposite a large number. It is because the number of communication pairs of l versus 1 required for service of n network nodes swells to n2. For example, when exchanging data between the data processing stations of n stand, the number of data-communications pairs of 1 to 1 is as follows. In ten node networks, the number of communication pairs is set to 45, and the number of communication pairs becomes 4950 from the upper formula in 100 node networks. Therefore, it is not practical to guarantee much reliability of communication of opposite a large number using the communication art of 1 to 1 at all, and it is insufficient even in a comparatively small network. Nevertheless, most prior arts use the communication pair of 1 to 1, in order to guarantee one pair of reliability of communication of opposite a large number in large numbers and in large numbers. Prior art is efficient and cannot provide the data communications of a large number to which reliability was guaranteed versus a large number. The problem of communication (broadcast and multicasting) of a large number to which the reliability by a digital-communications network was guaranteed versus a large number is pointed out by various persons. Some of examples of representation of the reliable report which has pointed out many problems of communication of opposite a large number are simply enumerated to below. Mai at al, 'Perf'oraance Evaluation of'anEthernet LANnuith Broadcast and Po1nt-to-Point Data Traffic", 10 Co Ll1 Puter CoIIlmuni-cation No, 6. Pp (Deeeo+ber 1987) 283-290 Waters, A, G The Use of'Broadcast andMultlcast Techniques on Coll1puter Net-vorks'', C0n1'erenCe On Electronic Delivery of Data and Software, IERE Publicati on No. 69, pp, and 45-50-(September 198B) 1 -- and -- this Reference Hughes and "Chat which are published at the document: an N-party talk facil ityf'or the Unlx 4.20perating 5ysteo+''. Coa+puter Communicat4ons (1988) Mr. Mai and others (Mat et at) is although a message can be essentially broadcast in an Ethernet type local area network, The problem of reliable broadcast has pointed out the point which has not been solved completely yet. The solution which Mr. Mai and others has proposed is providing a hierarchy in a receiving station and making it only a certain specific receiving station participate in actual broadcast or multicasting reception. Such a specific station (known as a rLSJ node) assumes duty to transmit the received message to other stations. In the Ethernet of the model of Mr. Mai and others, only the node as which - group was specified was allowed to send a broadcast message, and the sending node assumes the address of the specified rLsJ node to be what is known beforehand (a use is limited considerably in fact). LS node receives a broadcast message and generates advice of receipt. LS node is stored as a related mail message for groups to which the broadcast message is specified beforehand below. Since only LS node should transmit advice of receipt to a broadcast node, the number of advice of receipt decreases. However, this method has some faults. One of them is that there are restrictions that only a certain node can broadcast a message. In addition, furthermore it is said that a broadcast node must wait for the advice of receipt from the rLSJ node (communication process of 1 to 1 which has the number of recognition nodes judged in practice to be a small number which is a grade which a systems engineer can admit) of the total Why, there are also another restrictions. In order for many of conventional uses which a majority of some the reexamination and abstracts of communication art of opposite a large number are indicated, and are performed by the broadcast network to guarantee reliability to Mr. Waters (Waters)'s document, the point of using the communications protocol of 1 to 1 is pointed out. If solution of Mr. Waters is attained purpose-oriented by such an easy method that a difficult problem is surprised with broadcast art, he has come to a conclusion. To the conventional result with the relevance written in Mr. Waters's document, Project universe broadcasting Off 7 Silty's (Project U niverse Broadcast Facili-tles>) Banet (BANET), group communication (Group CoIImunication on the V-8ystem) on V system, Chang A broadcast protocol with the reliability developed by Mr and others (Chang et at), various file broadcast mechanisms, etc. are included. These are written in detail in the document quoted by the bibliography of Mr. Waters's document. In Mr. Waters's document, as for project universe broadcasting facilities (section 4.1), a "Cambridge ring (Cambrid ge rings) J type network is a broadcasting satellite channel. It is indicated that it is the system linked. It is arranged so that many numbers of broadcast channels may enable communication of opposite a large number. However, in this composition, true broadcast disappears at a satellite access station, and is transmitted to all the stations where the packet has the reception demand of the broadcast message of a specific channel beforehand in the local ring network by the access station. Therefore, it must be told beforehand which specific node must receive broadcast information, and the satellite access station of this system has only participated in the broadcast communication of truth [ station / satellite access ] actually. The Banet mechanism indicated into section 4.2 of Mr. Waters's document provides communication of reliable one-pair a large number by making each receiving station carry out advice of receipt of each packet. Therefore, when there are many receiving stations, it is not practical and is inefficient (if the member who has a receiving group especially fails in an answer, there is no telling what happens). The composition of V system (distributed processing system based on Ethernet) by Mr. Celi Don (Cheriton) and the Zwaino pole who stated to Mr. Waters's document previously is also indicated into section 4.3. Although this composition is due to communication of 1 to 1 to which reliability is guaranteed, Another message is sent after making an application program judge whether a sufficient number of replies (advice of receipt) were received, Some badness of the efficiency at the time of using communication of 1 to 1 for broadcast or multicasting is improved (some uses have few requirements about reliability than other uses). This composition is completely insufficient, when the broadcast which is reliable 100% among a huge number of nodes, and multicasting communication are required. The broadcast protocol which has the reliability developed by Chuck (Chang) and Mr. Mexamcuk (Mexa mchuk) to Mr. Waters is also indicated. The 2nd volume of the ACM transaction (A CM Transactionson Computer Systems) according [ this protocol ] to computer systems, No. 3 and 251-273 (August, 1984) "it is indicated by name called high-reliability broadcast protocol (Rel 1 ableBroadcast Protocols) J. Both broadcast art about transmission of the first message and transmission of advice of receipt of a message is used. The single site called a "token site" transmits advice of receipt to a broadcast message addresser. Other stations transmit only negative advice of receipt, and hear the message and advice of receipt by a token site. A token comes out via an advice-of-receipt message, and passes from a certain station to another station, and each station stores the packet which enters. Thus, the number of the advice of receipt transmitted for every message decreases. However, advice of receipt other than the message it was broadcast that every station was must always be heard. Finally, into section 5.1.3 of Mr. Waters's document, the file dispersion method developed by two persons (Sab nani) of Subna and Mr. Schwartz (Schwartz) is indicated. This uses the negative advice-of-receipt broadcast dispersion techniques developed by Goh Buck N alternative repeat art (go-ba ck-N 5 electlve repeattechniques), Mr. Callot (Ca l o), and Mr. Easton (Easton). The utility program which is called "chat (Chat) J for JEX operating systems to Mr. Fuse's (HugheS) document and which was run in fact is indicated. Although this utility program is having broadcast and multicasting of a message performed, Since it is not guaranteed at all, it is not who [ any ] other than datagram type solution whether the message is correctly transmitted to all the members of the multicasting set, either. This solution is insufficient for the use of almost all broadcasts and multicasting in respect of reliability. This invention described below solves many main problems in one-pair a large number or the data communications of data of data communications of opposite a large number of guaranteeing reliability. Some different special communication functions are made to distribute and carry out on a communication network in this invention by making a function which is different in a different network node carry out. Various data processing stations can be made to distribute and carry out various special functions in a certain example. Or some functions can be carried out by the same network node, and the number of data processing stations can be reduced. Or one special function - Existing can also be divided and carried out in some separate data processing stations. At a desirable example, they are these special functions (since each function is carried out by different network station in the desirable example, it calls the following "special station"), A means to guarantee that the station of the total Why on a network receives the message of the broadcast total Why correctly (with 100% of reliability [ Substantially ]) is provided. The following are contained at these special stations. (1) General subscription station (2) retransmission-of-message station (3) specification record station (4) record reproduction station (5) management station (arbitrary) The stations of the total Why connected to physical transmitter Seki and system (for example, share digital-communications network) can access the transmitter Seki and system all at once. General subscription station (one or more personal computers, workstation, and the data processing station of the special or general object are comprised), It functions as the source of dispatch of information, or an addressee of information, and decides whether to participate in communication or a meeting of which group based on the contents (for example, meeting number filed in the special protocol message header of the desirable example) of the message protocol. A retransmission-of-message station receives a human power message by 1 to 1 from a general subscription data processing station, adds a special protocol message header to an input messsage, and this message is turned to the data processing station of the total Why, and it broadcasts it again (broadcast). A receiving station judges whether each received the message correctly. When there is no error in the message received in the desirable example, A receiving data processing station passes a message (The). For example, it judges based on the actual information included [ whether it is Send and ] in the application program at the message header (for example, the meeting number field, the transmitting people discernment field, the message type field) or the message. According to another feature of this invention, it is the reliability of the data for the data communications of one-pair a large number or a large number versus a large number, Detection of the message lost or damaged, and it is guaranteed by introducing the new communications protocol to which all the data processing stations enable it to carry out reception of a message or recovery of each subscription data processing station to or -- the predetermined group of a subscription data processing station separately. In a desirable example, the lost message is detected by inspecting the message sequence number field of the header of a special communications protocol, and the damaged message is detected via the C RC field or other error-checking-and-correction fields. This special protocol of a desirable example comprises the time stamp field, the meeting number field, the transmitting people discernment field, and the other fields. The message total Why transmitted after the date which is a receiving station and was given by these fields, for example, or time can be recovered, or a message can be chosen based on a meeting number or transmitting people. A specification record station guarantees that the message sent from a transmitting station is exact. This specification record station carries out efficient communication on a retransmission-of-message station and 1 to 1 base via a broadcast network. Transmission of the message from a retransmission-of-message station to a specification record station is received by all the receiving stations. If it detects that a specification record station has an error and a loss in a message, it will be requested that a message should be broadcast again to a retransmission-of-message station. Since this specification record station also checks reception of the message from a retransmission-of-message station, in the desirable example, the message speed of a retransmission-of-message station can be adjusted by changing a time interval until it transmits a confirmation message. Exact reception of the message of the total Why to which a specification record station is transmitted by the retransmission-of-message station is guaranteed. The message received correctly is temporarily memorized by the specification record station. A reproduction record station receives and memorizes the message from a retransmission-of-message station. This station comprises one or more sets of data processing stations (when using one or more sets of data processing stations). Based on the contents and the meant addressee of a message, some stations receive alternatively the message of the total Why which has a certain meeting number field value, and memorize it somewhat permanently, for example. Detection of that a reproduction record station has an error and a loss in a message will judge whether it waited until it received at least one another message, and it produced depending on how to receive a message for whether the previous error arose at the retransmission-of-message station in the desirable example. When an error arises at a retransmission-of-message station, a specification record station must detect an error and must demand retransmission of message of a message. If the error arose by reception of the reproduction station, a reproduction record data processing station will demand the right information using the communication art of 1 to 1 in which reliability is guaranteed to the specification record station. As stated previously, the data transmission technology of 1 to 1 with which reliability is guaranteed is used for communication between a retransmission-of-message station and the specified record station. Therefore, the reproduction record data processing station of the total (the data exchange of 1 to 1 is heard) Why is guaranteed so that a meeting can be recorded correctly and completely. Since some reproduction record data processing stations are used in the same network, they can make one or more sets of reproduction record data processing stations able to memorize the same data, and can memorize a message redundantly. By this redundant memory, even if operation of a certain reproduction record stage gin stops, information is recoverable from another reproduction record stage gin. If general subscription data-processing stage gin detects the message disappeared or damaged, the right information can be soon acquired from a reproduction record station using the data transmission technology of 1 to 1 with which reliability is guaranteed. Therefore, even if the general subscription data processing station of the total Why does not transmit an advice-of-receipt message to a retransmission-of-message station soon, receiving the message of the total Why correctly from a retransmission-of-message station is guaranteed. In another mode of this invention, the reproduction record station also has a storage function. The general subscription data processing station can carry out the total Why request of the special meeting message after a predetermined date and time, for example. The general subscription data processing station can obtain again the meeting message of the total Why lost when the power supply was switched on from a reproduction record station, even if the power is turned off for a while. The management station which enables a regular start and stop of group communication or a meeting can also be provided. The management station can also deal with the held secret communication, and carries out data encryption for approval of a user, recognition of a meeting, detection of a network error, distribution of a cryptography key or a decipherment key, and a secret conference. The special station and function which are provided by the invention of 2 are distributed on a communication network, and communication of guaranteed very efficient one-pair a large number is realized. For example, the case where a general subscription data processing station transmits the same message to 1,000 sets of other data processing stations is considered. A general subscription station obtains use permission of a meeting number for example, from management stage gin first, and sends a message to a retransmission-of-message station using the data transmission technology of 1 to 1 with which reliability is guaranteed next. A retransmission-of-message station adds a special protocol header to a message, and broadcasts a message again. The message broadcast again to 1,000 sets of the meant receiving data processing stations is heard, and each processing station chooses a message based on the meeting number field. Each receiving station which was not able to receive the broadcast message well, Reception of data is guaranteed, even if specific receiving data-processing stage gin is processing or the switch is turned off, when a message is broadcast first since the retransmission of message from a reproduction record station is demanded. Thus, in the best case, data is only transmitted twice slightly. As for the 1st time, the 2nd time is to the station of other total Why from a retransmission-of-message station from a send data processing station to a retransmission-of-message station. (It is important to transmit data twice.) In almost all the prior arts about broadcast, it is going to secure reliability only by transmitting data once [ only ]. Many data communications of opposite a large number are simply attained only by [ with the sufficient efficiency in which reliability was guaranteed ] extending slightly the art of above-mentioned one-pair a large number. 1,000 sets of data processing stations -- which -- although -- it can write in the same (specified by the same meeting number) meeting using the art of above-mentioned one-pair a large number. 1,000 sets of data-processing stage tongues can communicate mutually, where reliability is guaranteed. Fundamental art can also be provided with the additional feature. For example, a safety feature can be provided using a management station. A meeting can be held secrecy and only the group selected from subscription data processing stations can read data using attestation and a code. Only a certain subscription data processing station can write in, and many of other subscription data processing stations also have the meeting where it is allowed to read. There are much one-pair large number to which reliability was guaranteed, and big influence about how a data processing station should be designed and used for the data communications of opposite a large number. In the use for which many are existing, although the data communications of opposite a large number are desirable in large numbers, since reliability is missing at one-pair a large number or the publicly known art about broadcast or multicasting, under the present circumstances, the inefficient communication art of 1 to 1 is used. Before developing useful application, reliability needs to be guaranteed first. Why does it attach importance to a guarantee of reliability in until so much? In a certain application, the absolute guarantee of reliance is clearly required. For example, even if it was [ of the hundreds edges ] small or several bits reception had a defect in distribution of software, the received software program is completely worthless. When imperfect software is used, there is a danger of destroying a data processing station or destroying the whole network. In order to show the effect of reliability, the artificer conducted some experiments. The datagram broadcast message was received from the IBM PC computer on Ethernet using one set of IBM P C/AT pair 21 Tha. That is, the message was broadcast, without the advice of receipt of a message also conducting the inspection of errors. The speed of IBM PC is slower than the speed of IBM PC/AT. An error ratio is private contract 1 of 100,000 messages which transmitted. This is a wonderful thing. Next, the datagram broadcast message was received from IBMP C/AT using IBM PC. Since IBM PC could not catch up with IBM PC/AT, 60% of messages were lost. If in other words early and a receiving side have the late transmitting side, even if it is the hardware which it is reliable and an error does not produce, a receiving side will lose a message. A certain station of users has a high possibility that capability is inferior compared with other stations, or capability is inferior to the mainframe or the mini computer in the environment of a typical actual life. The situation where the sending side broadcasts at a speed earlier than a late receiver may arise easily. It is rare that a message continues continuing by the message wastage rate actually exceeding 60%. This point is just one big reason that cannot use many practical Abridgoschin for broadcast if there is one pair of no data transmission technology of opposite a large number in large numbers or in large numbers reliability is guaranteed. The unsolved problem is left behind to realization of the communication of a large number versus 1, or a large number versus a large number with the sufficient efficiency in which reliability is guaranteed until it continues up to now. However, it is not necessarily said that such communication is unnecessary. In fact, if one pair of reliability of the data communications of opposite a large number can be guaranteed in large numbers or in large numbers, in a data processing station, many various uses will newly arise. A part of examples of a new use are enumerated below. (1) Distribution of stock market trading system (2) software (3) Teaching or demonstration (4) Communication with a strange receiving group (5) Network management (6) Network safe surveillance (7) computer conferencing (8) The support (9) backup system of a different company (10) Hot standby system (11) on-line-transaction processing (12) The distributed database (13) method humanity news system (14) One-pair a large number with processing fist control system (15) parallel phi processing type phi application (16) reliability, and other uses which receive much profits from the data communications of opposite a large number The paragraph of "the use of this invention" explains the above example in detail. Brief explanation of the drawings Although he could understand better these of the present invention, other features, and an advantage more completely by referring to detailed explanation of the following of a desirable example in relation with the attached drawing, ; figure 2 which is a chart diagram of the present desirable example of communication system S by which :figure 1 follows the present invention in here, ; figure 3 which is a chart diagram showing exchange of the typical communication in the system of Drawing 1, ; figure 4 which is example hardware constitutions of the structure of the example station shown in Drawing 1, ; figure 5 which is a chart diagram of the example message format of the communication protocol used by the example of Drawing 1, It is a detailed chart flow diagram of the initial-setting block; figure 6 which is execution I Eagram is indicated to be to Drawing 5 by the re-transmission station shown in Drawing 1; ; figure 8 which is a A batch in which Drawing 7 is shown in Drawing 5, and a detailed chart flow diagram of a weight loop block, ; figure 9 [ detailed / of processing routine block ACKTIM shown in Drawing 7 ] which is performed when a response timer becomes the due date and which is a chart flow diagram, ; figure 10 which is a detailed chart flow diagram of processing routine MERC2 shown in Drawing 7 performed after a message is received in network B, :figure 11 which is a detailed chart flow diagram of processing routine MRECI shown in Drawing 7 performed after a message is received in network A is a detailed flow diagram of the fatal error-handling block shown in Drawing 5; Drawing 12 is a specified recorder Sutee table flow diagram which is shown in Drawing 1; Drawing 13 is a detailed flow diagram of the initial-setting block shown in Drawing 12; ; figure 15 which is a detailed block diagram of the A batch in which Drawing 14 is shown in Drawing 12, and a weight loop box, When a message is received by the recorder station specified, it performs, Processing routine RE CMG shown in Drawing 14; figure 16 which is a detailed block diagram of two, One of the playback recorder stations on a network ; figure 17 which is a detailed flow diagram of processing routine RECMG1 shown in Drawing 14 performed when a message is received, ; figure 18 which is a detailed flow diagram of the fatal error-handling block shown in Drawing 12, ; figure 19 which is a chart flow diagram of the example programmed control step performed by the playback recorder station shown in Drawing 1 is a detailed flow diagram of the initial-setting block shown in Drawing 18; ; figure 21 which is a A batch in which Drawing 20 is shown in Drawing 18, and a detailed flow diagram of a weight loop block, When a message response timer serves as the due date in network A, a playback recorder station performs, ; figure 22 which is a detailed flow diagram of the ACKTMO processing routine block shown in Drawing 20, When a message is received in network B, a playback recorder station performs, ; figure 23 which is a detailed flow diagram of processing routine block MESON2 shown in Drawing 20, It is a detailed flow diagram of processing routine block ?vl E S ON1 shown in Drawing 20 performed by the playback recorder station when a message is received in network A, and is Ah. To: Drawing 24 is a detailed flow diagram of the fatal error-handling block shown in Drawing 18; It is a chart flow diagram of the example programmed control step Execution(ed); ; figure 27 which is a flow diagram of the initial-setting block Drawing 26 is indicated to be to Drawing 25, ; figure 28 which is a detailed flow diagram of the dispatch shown in Drawing 25, or a weight loop block, When a re-transmission computer 1 response timer becomes the due date, a general participating station performs, ; figure 29 which is a detailed flow diagram of processing routine ACKTMI shown in Drawing 27, ; figure 30 which is a detailed flow diagram of processing routine ACKTM2 shown in Drawing 27 performed by the general participating station when a response timer becomes the due date, When a message is received in network B, it is a detailed flow diagram of processing routine M S G ON 2 shown in Drawing 27 performed by the general participating station; Drawing 31 is performed when a message is received by the general participating station in network A, ; figure 32 which is a detailed flow diagram of processing routine block MSGON I shown in Drawing 27, It is a detailed flow diagram of processing routine block KEYINF shown in Drawing 27 performed when inputted via the keyboard of a general participating station as the Full message which should be transmitted shown in Drawing 1 at, or other input devices; Drawing 33 is a detailed flow diagram of the fatal error-handling block shown in Drawing 25: :figure 40 whose Drawing 34-39 is the chart diagram by which still the example example of the present invention in which it did not limit was generalized shows the example data structure about a re-transmission computer; ; figure 42 Drawing 41 indicates an example additional data structure about a re-transmission computer to be, A re-transmission computer on network B which is set as Drawing 40, And; figure 43 which shows the example message control block used in order to pursue the work between the recorder computers specified, A re-transmission computer on network A which is set up in Drawing 40, And; figure 44 which shows the example message control block used in order to pursue the work between common participating computers shows the example data structure about the recorder computer specified; Drawing 45 shows the example additional data structure about the recorder computer specified; A recorder computer on network B by which Drawing 46 is specified in Drawing 44 and which is specified, And; figure 47 which shows the example message control block used in order to pursue the work between re-transmission computers, The recorder computer on network A which is pinpointed in Drawing 44 specified, And; figure 48 which shows the example message control block used in order to pursue the work between playback recorder computers shows the example data structure about playback recorder convenience store -Tar; ; figure 50 Drawing 49 indicates the example additional data structure about a playback recorder computer to be, A playback recorder computer on network B which is specified as Drawing 48, And; figure 51 which shows the example message control block used in order to pursue the work between re-transmission computers, A playback recorder computer on network A which is specified as Drawing 48, And; figure 52 which shows the example message control block used in order to pursue the work between the recorder computers specified, A playback recorder computer on network A which is specified in Drawing 48, And; figure 53 which shows the example message control block used in order to pursue the work between general participating computers, ; which shows the example data structure about a general participating computer -- general participating pair Nita on network B by which; figure 55 Drawing 54 indicates the example additional data structure about a general participating computer to be is specified as Drawing 53, And :figure 56 showing the example message control block used in order to pursue the work between re-transmission convenience store -Tar, General participating pair Nita on network A which is specified in Drawing 53, And; figure 57 which shows the example message control block used in order to pursue the work between re-transmission computers, A general participating computer on network A which is specified in Drawing 53, And; figure 58 which shows the example message control block used in order to pursue the work between playback recorder computers, It is the conventional system for; figure 60 which shows the example message control block about all the stations where; figure 59 which shows the example message block about all the stations shown in Drawing 1 is shown in Drawing 1 in which the mistake was made to process the stock quotations in a securities firm; ; figure 62 which is an example system for Drawing 61 to process the stock quotations by the present invention is an example system for processing the software distribution by the present invention; Drawing 63 shows the example system for performing the education by the present invention, and a demonstration; ; figure 65 which is an example system for Drawing 64 to process communication with the strange receiving person group by the present invention, ; figure 66 which is an example system for processing the network management function by the present invention is an example system for processing the network preservation monitor by the present invention; Drawing 67 shows the example system for processing the convenience store -Tar meeting by the present invention; Drawing 68 shows the example system for processing the support for a different company by the present invention; ; figure 70 which is an example system for Drawing 69 to process backup of the data processing station by the present invention is an example system for processing the urgent standby by the present invention; Drawing 71 shows the conventional example system for treating on-line-processing application; Drawing 72 shows the example system for treating the on-line-processing application by the present invention; Drawing 73 shows the example system for processing the distributed database by the present invention; Drawing 74 shows the example system for processing the business information system by the present invention; Drawing 75 shows the example system for processing the process control system by the present invention; ; figure 77 which is a useless example system which processes a patent search according [ Drawing 76 ] to the present invention, ; figure 78 which is a generalization chart diagram of the 7th example of the present invention, ; figure 79 which is an example system for performing the 7th example of the present invention, ; figure 80 which is a generalization chart diagram of the 8th example of the present invention, ; figure 81 which is the example format of the response message examined in the 8th example of the present invention, It is chart Diamondrum showing exchange of the typical communication in the system of Drawing 79, and a certain :figure 82 is example composition which shows the simulated broadcast in an extensive field network environment. In order that detailed explanation of a desirable example example may promote a quicker and easier understanding of the concept of the present invention by a person skilled in the art in the detailed explanation of a desirable example example present in the present, it is constituted as follows. In the present of communication system S when 1st the present invention is followed, the structure of a desirable example example is described about Drawing 1. The overall method with which the example of Drawing 1 communicates a message is described in connection with Drawing 2 now. And the hardware constitutions of various stations are described in relation with Drawing 3. Next, the example message protocol used by the desirable example is described in relation with Drawing 4. After that, the operation of the example of Drawing 1 is considered in relation with the chart display of the flow chart of the example programmed control step shown in Drawing 5-33 and 40-59, and an example data structure. Alternative network structure is supplied and the additional example of the present invention of combining an abundant function in/or the same network station is examined in relation with Drawing 34-39 after that. The delay in transmission of :, example error recovery technical;, and the present invention which examines some additional features of the present invention which includes the following after description of various kinds of alternative examples is an important factor, Example art for broadcast capability to use example technical; and the present invention for using it in health communication in the network environment (getting it blocked token ring network) which is not clear; - Example art which can carry out the ink face of example I\-Dwarf arrangement; and - application programmer 1, or the end user which may improve the efficiency of a system to the communication system provided by the present invention. - Example directions of the present invention (it describes in relation to Drawing 60-76). 1st example figure 1 provided by the present invention is a chart block diagram of a desirable example example in the present of communication system S. Communication system S is carried out in network A16 and 2 bus network composition containing 818. This example shown in Drawing 1, The;(a) re-transfer function provided with the following system functions (re-transmission station 20) The recorder function by which; (b) specification is performed and carried out (recorder station 28 specified) It performs; (C) playback recorder function (M performs from playback recorder station 1 in 26);, and (d) general participating function (N performs from general participating station 1 in 24). Each main functions provided by system S according to the present invention in the example of Drawing 1, One or the separated network station beyond it (each of these comprises the individual computer in a desirable example so that it may explain henceforth) performs. It is because the number of 1 meetings is assigned a priori so that it may explain after - and only the meeting which is not secret is supported, because the controlling function or the station is not included in the example of Drawing 1. In order to reduce the necessity of inventing as clearly as possible and conducting many experiments, the specific hardware about the 1st example and operation system software are explained here. However, the present invention should take notice of the thing which have what kind of operation system and which will be * Said if almost can be carried out also to what type of computer. The only necessary information is only having to support the communication interface which can access the communication equipment with which such a computer is shared. A token ring, a bait- network (Ethernet), radio, microwave, and a health network are included in the example of such a thing. In the 1st example, the both sides of :(a) network A16 whose hardware used is as follows, and network B18 are bait- networks (E thernet). (b) All the computers (re-transmission computer 20, specification recorder computer 28, playback recorder 26, and general participating computer 24), It is mjero VAX 11 computer made from DigitaI Equipment Corporation. (C) A bait- Internet communication interface is DEQNA of the same Digi tal Equipment. (d) Operation system software is VAX/VMS. : (a) whose reason for the above-mentioned selection is as follows It became clear that Micro VAX I I was reliable. (b) One set of Micro VAX II can support two sets of DEQNA:7 Municipality interfaces. (c) A V A X /V M S operation system provides the directions of the system service to the following things, and transmits and receives a message with : and the DEQNA communication interface which fully document-izes this; -; which sets a timer -- process - keyboard interruption. (d) VAX/VMS is multiuser required for the test of other examples (2-8), and a multi-work operation system. The number of the computers used by giving many functions to a single computer according to these examples can be reduced. (e) V A X /V M S provides the outstanding program development tool. (f) The program which VAX/VMS was already able to grant the privilege can make it enable it to choose a message based on a protocol type or a group (multicasting) address. (g) VAX/VMS always has communications skills of good l versus 1 in D E Cnet. Coexistence (in this case, DE C net) with the network software besides :- the following things can be proved [ network software ] if the present invention is carried out by VAX/VMS; - as opposed to other network software (in this case, DE Cnet) -- one pair -- a large number -- or the improvement with capability extensive in large numbers about application of opposite a large number. (h) In order to keep hardware possible [ simplicity and exchange ], they are all the Micro V. A X II must have the same composition. Jam: - MicroVAX processor;, 4 M bytes of memory; two -DEQNA interfaces; - Hard disk 2 sheet;, one or more terminals in which each can store 100 M bytes or more of data; - The device (it is product made by Di gital Equipment CorpOratIOn to the appearance) present invention which loads operation system software like TK50, It should be cautious of the ability to carry out also to other makers' (for example, IBM, an apple, Wang1 Unisys, Sun, etc.) computer. In fact, the present invention is extremely suitable for the communication in the environment which mixed the maker. : (a) by which or is simplified a little in the functionality in that the example shown in Drawing 1 is based on assumption or limitation (it will probably be unsuitable about a certain thing although it is suitable for many examples of application) of the following [ this ] -- all meeting is specified a priori and known by all participants. Up to meetings 1-10, it is saved for system use and can be used for meeting No and the meeting which is not secret as for 10,000 or more; (b) At the time of generating of the;(C) fatal error by which only the meeting which is not a secret is supported The whole system S stops immediately and;(d) memory buffer disk store or the system resource error like disk reading / write error which waits for processing by an operator does not have (error of these). It can process to a person skilled in the art using the publicly known art of all the <former; The communication art of 1 to 1 by which (e) use is carried out is the easy former type communication art in which a person skilled in the art is publicly known. Two separated networks in the example of system S shown in Drawing 1, certain : -- they are network A16 and network 818. Station 20.24 and all of 26.28 have two separated communication interfaces by which one is connected to network A16 and another side is connected to network 818. Network A16 is used for communication of 1 to 1, and inconsistency is solved using the conventional network technology (for example, use of the conventional Ethernet protocol), Network B18 shown in supporting-simultaneously-many pairs of communication of 1 to 1 figure 1 is a special network in a desirable example used for broadcast messages. The only pairs of the network station which communicates mutually via network 818 (and it has the capability for each to transmit a message) are re-transmission station 20. and recorder station 28 specified. Other stations only hear altogether this single pair of station 20.28 which talks mutually via network B. Network technology needed for this network 818, Between far easier 1 re-transmission station 20 than what is needed for network A16, and recorder stations 28 specified, it is because there is only one pair of communication of 1 to 1 (and since it comes out, the inconsistency with a mutual station cannot take place). It may be thought that all other stations only monitor the conversation between these two stations. In this operation, a standard communication interface can be used for re-transmission station 20, the recorder specified, and stations 28. The communication interface for other computers of all the is further easier. These do not have the necessity of consisting transmission logic. - It is merely only receiving logic. In the operation in actual everyday life, it is possible to be excessive, and can make selection of having the kind same about the both sides of network A16 and network 818 of network. That is, when network A16 breaks down, network 818 can play a role of backup and is possible also for the contrary. Communication of guaranteed reliable one-pair a large number can be attained as follows: (a) Concerning Drawing 1, general participating station 24 (1) considers that the same message would like to send to general participating station 24(2)-24 (N), for example. General participating station 24 Using 1 reliable communication art versus 1 in which network A16 was guaranteed, (1) sends a message to re-transmission station 20 first. General participating station 24 (1) specifies similarly that meeting No and 10,000 should be used. (b) Re-transmission station 20 changes a message into a special format as shown in Drawing 4. Re-transmission station 20 re-transmits a message to recorder station 28 specified on network 818 from it. Since all other stations 24.26 are monitoring the communication on network B18 similarly, re-transmission station 20 will be broadcast to all the stations at the same time it transmits it to recorder station 28 which has a message specified substantially. Recorder station 28 specified checks that the transmitted message is exact by returning the message which accepts clear reception of a message to re-transmission station 20 via network B18. This carries out the short store of the message received certainly similarly (for example, less than 10 minutes). As for Recording y- and station 28 specified since there are the message cognition and the error detection mechanism in a desirable example, receiving correctly all the messages from re-transmission station 20 is guaranteed (probability of not detecting 1O-13 of a theoretical error). On the other hand, since these are physically connected to network B18, all the stations are Listen (.) about communication between re-transmission station 20 and recorder station 28 specified. Each of these stations can detect whether this made the mistake in the message by inspecting the message sequence number field of special message header 44 (refer to Drawing 4). For example, supposing a certain station receives message 1, 2.3, and 4*6, it will get to know that message 5 could not be received. Thus, a lack message is always detectable by inspecting message sequence number field 48. Although it may, of course, say that a message cannot be altered, cannot be distorted or is not received at all, Although such an error is detected by the art of the Cyclic Redundancy Check (C RC) known well, it is necessary to acquire the C RC value which inspects the data in which a transmission data processing station should be transmitted in this case, and is added to the message which should be transmitted. If Drawing 4 (chart block diagram of the example special message protocol about the message transmitted by re-transmission station 20 according to the example example of the present invention desirable now) is described briefly here, Three portions in all the messages transmitted by the re-transmission station in a desirable example, It is contained: (1) -- network header field 42(contents determined by specific type network B currently used):(2) special message header 44(generated by re-transmission station 20): -- and (3) error checking / reform field 46 * -- if it says simply -- special message header 44 :(i) message sequence number field 48 (assigned by the re-transmission station according to a sequence) where the following fields are included (refer to Drawing 4) ;(i i) stamp field 50 (a re-transmission station) to which a message is transmitted by the re-transmission station Time:(iii) meeting number field 52 (setting in the desirable example) to which a message is transmitted in first stage Discernment supplied by general participating station 24 whose re-transmission station is a starting point of a message, And this field value is assigned according to/or meeting number information, and this field value is used by the general participating station of a receiving side in order that they may determine whether be a receiving person by whom that message is meant; (iv) An informer's discernment field 54 (this field) The identifier of - mind relevant to general participating station 24 which is a starting point of a message is contained : A (V) message type, field 56;hi error code field 58;(vii) protocol version <> field 60; (vili) Retry counter field 62 (in order for the present transmitted message to be denial by recorder station 28 specified or to show the number of times of having been repeated according to the negative response) ; (1x) preservation field 64 used by re-transmission station 20; it reaches. (X) Data field 66 (the contents of this field are generated by the general participating station which serves as a starting point typically, and constitute the useful message data which should communicate). After reception of a message, using the same CRC art, a receiving station inspects data and acquires a CRC value. When this CRC value is the same as what was sent by the transmission station, this message is accepted noting that it is right. When this C RC value differs from what was sent by the transmission station, this message is treated as the thing which was able to be distorted, or a thing which consists a CRC error. General participating station 24 is always detectable using the error detection art of the above-mentioned explanation about whether for the message received from re-transmission station 20 to be perfect, or to be exact. When general participating station 24 detects an error, this makes the information righter than playback recorder station 26 profitably like. Note that there is no message response to re-transmission station 20 from general participating station 24. in this way, the message sent to recorder station 28 which general participating computer 24 disregards all the messages other than meeting No and the thing of 10,000, and is specified from re-transmission station 20 when errorless -- Listen (being sufficient .) As for this, I hear that general participating computer 24(2)-24 (N) can receive all the messages on the basis of general participating station 24 (1) altogether. When an error occurs, playback recorder station 26 involves. (C) Playback recorder station 26 plays many roles like general participating station 24. Similarly; and this which hear the message sent to recorder station 28 where playback recorder station 26 is specified from re-transmission station 20 via network B do not return a message response to re-transmission station 20. However, playback recorder station 26 has the responsibility which records all the messages which belong to one or the meeting beyond it by extensive store means 26A. The recorded message can be held for several months or several years several weeks several days. A store means may be a means of a disk, a tape, or others. New write once lead minium Thailand Mus (WORM) laser technique can also be used. In order to record a specific meeting, it is desirable to use one or more playback recorder stations 26. Even when carrying out like this and use of playback recorder station 26 which records meeting 10,000 becomes impossible, information is more nearly available than another playback recorder station. various actual record art used comes out at playback recorder station 26. Continuous record and search are only one of much art considered. The actual record art of playback recorder station 26 which can function similarly as a file server for many general participating computers 24, and is used may be specific to a file server. When it is detected that a lack message is one of the messages to recorder station 28 where playback recorder station 26 is specified from re-transmission station 20, This recovers the message which used 1 to 1 pair of communication art in which Faith on network A16 guaranteed ("C") could be set, and was missing from recorder station 28 specified. When playback recorder station 26 detects the message (or message which has a CRC error) which was able to be distorted, this does not record the message which was able to be distorted. Instead, the following message is dealt with like, although waiting and its message which was able to be distorted were lost. (d) General participating station 24 can make reception of all the messages from re-transmission station 20 guarantee. the message missing from reliable playback recorder station 26 where this was guaranteed using the reliable art of 1 to 1 guaranteed via network A16 when general participating station 24 detected an error -- To -- things are made. General participating station 24 can cut a switch for many days or many weeks. when a switch is returned again, this transmits all the lack messages of a period that had turned off the switch to playback recorder station 26 -- as -- To -- things are made. This is specially made as [ show / in Drawing 4 ] because [ of sequence number field 48 in message header 44, and time stamp field 50. ]. Using time stamp 50 and message sequence number 48 information, the software on general participating station 24 records the message received correctly at the end, before the switch of general participating station 24 is turned off. When again put into the switch of general participating station 24, this is To in like [ slight / which recovers all the missing messages before asking first the 1st on network B18 again at playback recorder station 26 ]. if it carries out like this -- general participating station 24 (two -- > -24 -- (-- N --) -- a meeting -- 10,000 -- being related -- general -- intervention -- a station -- 24 -- data communication of reliable one-pair a large number which were ensured [ receiving the message from (1) and ] and were guaranteed is attained.) The thing which change the schema of the above-mentioned explanation and can perform guaranteed reliance and for which many data communications of opposite a large number are attained is very easy. General participating station 24 is a meeting altogether! It can write in 10.000. Informer identifier field 54 for identifying an actual informer is included in special message header 44 as shown in Drawing 4. Thus, general participating station 24 (the message about meeting 10.000 is sent and it is [/or ] winning popularity) can communicate between oneselves, or can attain many data communications of opposite a large number. The guaranteed data communication of a large number versus reliable a large number is attained. Detailed explanatory views 40-43 of a re-transmission station show the data structure used by the re-transmission station. A data structure is like the message received on numerical;(b) CRC error count 822 which is not changed dynamically as shown in :(a) component name 802 arranged as follows and node address table 804, and network A flag 858. The pointer to other data structures like numerical;(C) message control block 834 changed as the program operates, and message block 913. A message control block is a data structure used in order to manage the traffic of the message between a pair of stations about a certain specific meeting. For example, message control block 834 manages the message traffic between re-transmission station 20 and recorder station 28 specified. In this case, meeting number 884 has no meaning. About message control block 836, meeting number 884 contains number 10,000. another meeting like meeting 10200 -- the same -- re-transmission station 20 and general participating station 24 the case where it exists between (1) --; -- in order to manage this meeting 10200, an individual message control block is used. A message block includes an actual message and the pointer to the next message block. Drawing 58 shows the actual format of a message block. Note that the data structure is designed so that in common about all various kinds of stations. As a result, if not used, the field may be in the specific data structure which are marked by carrying out. Drawing 5-11 is a flow chart of the exemplary programmed control step performed by re-transmission station 20 shown in Drawing 1. The overall system logic flow about a re-transmission station is shown in Drawing 5. If re-transmission station 20 is started, a temporary setup code performed in initial-setting 70 portion of processing will appear. However, main processings are performed in the A batch of processing, or weight loop 72 portion. When there is no fatal error, the A batch of processing or weight loop 72 portion is repeated repeatedly. Re-transmission station 20 is waiting (MSG received on network A flag 858) about a certain event, such as reception of a message, happening. Or and MSG received on network B flag 860 is set up, in order to process the event, a A batch is carried out to a special processing routine. After a special processing routine finishes, in control returning to the main lines of processing again and being errorless, re-transmission station 20 waits for the following event as shown in Drawing 5. When there is a fatal error, the bus of the control is carried out to fatal error-handling 76 portion for processing. After fatal error handling in the example this program is indicated to be exists. Drawing 6 shows initial-setting 70 portion of processing in detail. In order to show that it is only one of processing boxes as this shown in Drawing 5 in, note that one box is drawn on the surroundings of the whole processing. this processing is very easy -- 1 -- it is clear or is read in a certain early numerical value about a data structure. The message control block for : and DRC834 which this initial-setting process comprises from the following matters, The message control block for GPC1836, the message control block for GPC2838, It is truly about; and the completion flag 904 of transmitting preparation which sets to 0 all the counters in; and the data structure which sets up Messe for ... and GPCn 840, ; which sets other flags of all the to a sake -- all pointers other than the pointer to - message control Brock Liszt 832 are set to zero. Drawing 7 shows the A batch of Drawing 5, or weight loop 72 portion in detail. The event which should be processed by re-transmission station 20 is as follows. It is: (1) A response timer is used for re-transmission station 20 in order to test about the message which was lost or was able to be distorted. Re-transmission station 20 starts a timer, after transmitting a message on network B18. Usually, it is expected that the response message from recorder station 28 specified arrives before a timer becomes the due date. When a message is able to be lost or it is able to distort, recorder station 28 specified does not send a response message. A timer serves as the due date (the ACK timer which became the due date on network B flag 866 is set), and message 913 by which the last was not answered is sent again (noting that it is tested by determination Brock 88). This message can be taken up using the pointer to message 902 sent to the last in message control Brock for DRC834. The processing routine which sends a message is called ACKT 1M90. (2) Determination Brock 92 tests about the message received on network B18. Generally, this means that recorder station 28 specified has sent the response message. The processing routine used in order to process the received message is called MREC294. (3) Determination Brock 96 tests about the message received on network A16 (generally). What it is going to communicate with re-transmission station 20 is shown for the To object so that general participating station 24 may broadcast a new message on network 818 to re-transmission station 20. The processing routine for processing the message received on network A16 is called MRECI 98. I would like to be cautious of processing the due date of a response timer first, and a determination processing step receiving the message on network B18 continuously, and receiving the message on network A16 after that. - This is performed relatively to those relative priorities on actual living conditions. Drawing 8 -- processing routine ACKT -- 1M90 is shown. This processing routine is inputted when a response timer becomes the due date (generally recorder station 28 specified shows that a response message could not be returned to re-transmission station 20). This failure will be for according to CRC or the lack message of either re-transmission station 20 or recorder station 28 specified. Although it tries in order that re-transmission station 20 may re-(it can set to Brock 120) broadcast the message sent before, there is restriction about how many times it is tried. If this restriction (to the utmost re-trial-line restrictions 854) is exceeded (it is tested by Brock 114,116 like), it will be considered that this error is fatal. This happens, when recorder 28 specified breaks down or there is an error of unrecoverable hardware. In a fatal error, a fatal error indicator (fatal Ella most item 824) is set (Block 18), and re-broadcast is no longer carried out to it. Drawing 9 shows processing routine MREC294. This processing routine is inputted when a message is received on network 818. Generally, this means having arrived from recorder station 28 where a response message is specified. CRC error handling is very easy. The CRC error is memorized, and a message is thrown away completely and disregarded (Brock 134,136). The message accompanied by a CRC error will assume that it is a response message from recorder station 28 specified. If this message is disregarded, as inquired by processing routine ACKTIM90, a response timer will serve as the due date and, as for re-transmission station 20, it will send out a non-response message again. If a response message is received from recorder station 28 specified, the preparation which re-transmission station 20 broadcasts that the following message is for [ of the general participating stations 24 ] one will be completed. as for the message from these general participating stations 24, these are received -- Is called -- while waiting to be processed or broadcast, it is put into things by the waiting sequence. It is pointed out to this waiting sequence by the pointer to message Liszt 896 in message control Brock for CHD834 who was not seen off. When there is no message into the waiting sequence which waits to be processed and broadcast, (it being tested by determination Brock 138) and re-transmission station 20 are, When re-transmission station 20 receives the following message on network A16 and does not need to be registered into a waiting sequence (Brock 146), Immediately, the message from general participating station 24 only sets the completion flag 904 of transmitting preparation, in order to show that it can send. A response timer is canceled when a response message enters (Brock 148). When a message is in the waiting sequence which is waiting to be processed and broadcast, re-transmission station 20 broadcasts the next message of (Brock 140-144) in a waiting sequence. The message broadcast just now is taken out from a waiting sequence, and, as for; retry transmitting count 892 set to an imitation, the completion flag 904 of :transmitting preparation which is rebooted as for; response timer to which it is pointed out by the pointer to message 902 sent at the end is re-set to zero, The decrease part of the non-outgoing message count 894 is carried out only 1. Drawing 10 shows processing routine MREC198. This processing routine is inputted when there is a message received on network A16. This generally means that general participating station 24 is trying to send a message to re-broadcast station 20 for broadcast. CRC error handling is very easy. A CRC error is recorded, and a message is thrown away completely and disregarded (Brock 164*166). A message will assume that it is the demand which transmits from general participating station 24 (tested by determination Brock 168). When a message is disregarded, in order to show that re-transmission station 20 accepted the demand, there is no response message returned to general participating station 24. General participating station 24 sends the demand once again. Re-transmission station 20 can receive the same demand from general participating station 24 on quantity once (tested by Brock 174,176). Re-transmission station 20 is using this, and when a response message cannot be sent within the due date period of a response timer, it may happen. It sends the same demand again, assuming general participating station 24 that the demand back to front [ the ] was lost. However, this fact is noted down by retry count field 62 in special message header 44. In the 1 to 1 communication between re-transmission station 20 on network A16, and general participating station 24, a message can deviate from a sequence (tested by determination Brock 178). Since out of order [ general participating station 24 / any times ], this may happen. When this is rebooted, this may have lost the information about the message sequence of the last exchanged for re-transmission station 20. Since it is a message response scheme of 1 to 1, a message is acceptable in this situation from a sequence. In being errorless, re-transmission station 20 returns a response message to general participating station 24 (Brock 180), and tries to assemble and send a broadcast message (Brock 190). The message can be registered into a waiting sequence when the message on network B18 cannot be sent immediately. As it explained in processing routine MREC294, whenever a response message is received from recorder station 28 on network B18 specified, a message is pulled out from this waiting sequence every. Drawing 11 shows fatal error handling routine 76 about re-transmission station 20. The fatal error considered here is failing to receive a response message from recorder 28 specified. This may be caused by the problem of the hardware in connection with the message reception in the problem or re-transmission station 20 in recorder station 28 specified. Processing of a fatal error is very easy here. As for me, saying [ that a message is displayed (Brock 202) and a problem leaves after that ] hears that the whole meeting stops. Explanatory view 44-47 of recorder station 28 specified shows the data structure used by recorder station 28 specified. Drawings 44 and 45 show main data structures. Drawing 46-47 illustrates these data structures in detail. These data structures are similar to the thing about re-transmission computer 20 in respect of a format. Drawing 12-17 describes the detailed operation of recorder computer 28 specified. The message sent from :(1) re-transmission station 20 whose main functions of recorder station 28 in a desirable example specified are as follows ensures being received continuously and that these messages cannot distort. This is attained by sending a response message to re-transmission station 20 after receiving a message better than re-transmission station 20. A response message is not sent when a CRC error is included in a message. Re-transmission station 20 is expected to resend the message after the deadline period of a response that Drawing 8 is examined. The message which deviated from the sequence is treated as a fatal software error, and a program is withdrawn. (2) In being required, it stores a message temporarily so that playback recorder station 26 can be used. However, it is To in like [ slight / these send the right message to recorder station 28 specified when this detects an error ]. Drawing 12 shows the logic flow of the overall system about recorder station 28 specified. Starting of recorder station 28 specified will produce a setup routine temporarily which is performed in initial-setting 212 portion of processing. However, main processings are performed in the A batch of processing, or weight loop 214 portion. When there is no fatal error, the A batch of processing or weight loop 214 portion is repeated repeatedly. Control is passed by error-handling 218 portion fatal for processing when there is a fatal error. The after-processing program of a fatal error exists. Drawing 13 shows initial-setting 212 portion of processing in detail. Message control Brock for : and RTC1034 whom this processing comprises from the following steps, and; which sets up message control Brock Liszt who comprises message control Brock for PRC1036 --; which sets all the counters in - data structure as zero -- the completion flag 1104 of - transmitting preparation is set up truly. Other flags of all the set all the pointers other than the pointer to; and message control Brock Liszt 1032 set as an imitation as zero. Drawing 14 shows the A batch of Drawing 12, or weight loop 214 portion in detail. The event which should be processed by recorder station 28 specified, On 1 type target received on network B18 (Brock 230), as for :(1) message as follows, the processing routine called RECMG2 by re-transmission station 20 Tseczek whether a message contains a CRC error. A response message is sent when errorless. (2) A message is received on network A16 (Brock 234). Generally, this means that the request message was received from one of playback recorders 26, in order to search the lost message. The processing routine which processes this type of message is called RECMol 236. Drawing 15 -- processing routine RECMG2 2 B-2 is shown. This processing routine is started by recorder station 28 specified when there is a message received on network B18 (determination Brock 230 of Drawing 14). Generally, this means that the message arrived from re-transmission station 20. CRC error handling is very easy. A CRC error is recorded, and a message is thrown away completely and disregarded (Brock 244,246). This message will assume that it is a message from re-transmission station 20. When this message is disregarded, a response timer which was examined in processing routine ACK TrM90 on re-transmission station 20 serves as the due date. If it does so, retry count 62 will be in the state where the increment only of 1 was carried out, and a re-transmission station will resend the message. It must never generate and considers that the message which deviated from the sequence is fatal (Brock 248,250). The whole system must be closed in a desirable example. When an error is not detected, a response message is returned to re-transmission station 20 (Brock 252). Drawing 16 -- processing routine RECMCI -- 236 are shown. This processing routine is started when there is a message received on network A16 (determination Brock 2341 figure 14). the message by which playback recorder station 26 was generally lost as for this -- Doing -- things are meant. CRC error handling is very easy. A CRC error is memorized, and a message is thrown away completely and disregarded (Brock 264,266). The message will assume that it is a request message from playback recorder station 26. if this message is disregarded (Brock 268,270), a response timer (processing routine ACKTMO312) will be examined -- as -- it becomes the due date and a request message is sent again. When errorless, a message which is specified as a request message and which was lost is searched from a memory buffer or temporary memory storage, and is sent to playback recorder station 26 which is demanding (Brock 276). Drawing 17 shows fatal error handling routine 218 about recorder station 28 specified. The fatal error considered here is the message which deviated from the sequence from re-transmission station 20. Probably this will be main and will serve as a fatal software bug. Fatal error handling here is very easy. A message is displayed and the program from it withdraws. As for me, saying hears that the whole meeting stops (Brock 282). Explanatory view 48-52 of playback recorder convenience store -Tar 26 shows the data structure used by playback recorder computer. Drawings 48 and 49 show main data structures. Drawing 50-52 expresses a data structure in detail. These data structures are the points of a format and are similar to the thing about re-transmission convenience store -Tar 20. Drawing 18-24 expresses the detailed operation of playback recorder station 26. The Also with main following functions of a playback recorder station From :(1) re-transmission station 20 which is of, the message sent to recorder station 28 specified on network 818 is heard, and it stores from it based on a meeting number for the search after calling the message at general participating station 24. (2) When playback recorder station 26 cannot obtain an exact message by listening on network B18, this obtains a message more exact than recorder station 28 specified. (3) General participating station 24 usually obtains a message directly from network B18 on a target. However, the time of the standup of a station, or when playback recorder station 26 detects a message error, it may need to send one or the message beyond it. Drawing 18 shows the flow of the overall system logic about playback recorder station 26. When playback recorder station 26 is started, this performs a temporary setup routine in initial-setting 292 portion of processing. However, main processings are performed in the A batch of processing, or weight loop 294 portion. When there is a fatal error, the A batch of processing or weight loop 294 portion is repeated repeatedly. When there is a fatal error, control is passed by fatal error-handling 298 portion for processing. A program is withdrawn after fatal error handling. Drawing 19 shows initial-setting 292 portion of processing in detail. Message control Brock for : and RTC1234 who consists of a step of the following [ processing ], message control Brock for DRC1236, GPC, and message control Brock for 11238 -- message control Brock Liszt who consists of message control Brock for ... and GPC n 1240 is set up. Drawing 20 shows the A batch of Drawing 20, or weight loop 294 portion in detail. In a desirable example, :(1) response timer whose event processed by playback recorder station 26 is as follows becomes the due date on network A16 (determination Brock 310). From recorder station 28 where playback recorder station 26 is specified, a response timer is started, when requiring the lost message. That a timer becomes the due date means that a response was not received by playback recorder station 26. This may be based on a CRC error, disappearance of a message, the problem of the recorder station specified, or failure of hardware. The processing routine which processes a response timer is called ACKTM O312. (2) A message is received on network 818 (determination Brock 314). Generally, this means that the message was received from re-transmission station 20. This message has it confirmed by the routine called MESON2 whether it is errorless (Brock 316). When errorless, meeting number field 52 is inspected in order to see the store of this message whether carried out or not. In this example, only the message belonging to meeting 4Jk10.000 is stored. A message is stored using the structure of a linked list. Two pointers are used in order to pursue message Liszt. These are the pointers to message Brock 1230 received at the To pointer for A, and the last which points out Liszt's concentration at received message Liszt 1228 who points out Liszt's start. This processing routine is called M ESON2. (3) A message is received on network A16 (determination Brock 318). This thing; which CF& request message entered from general participating station 24 -- or (b) Mean having entered from recorder station 28 where a response message is specified. The processing routine called MESONI 320 processes the message from network A. Drawing 21 shows processing routine ACKTMO312. After being sent to recorder station 28 where a request message is specified, this processing routine is started when a response message does not enter from this (310 of determination block diagram 20). Unless the maximum retry count is exceeded (tested by Brock 328), this request message is resent (Brock 330). When the retry count has been exceeded, this error is forsaken as it is fatal, and is processed by Brock 332. Drawing 22 -- processing Rotin MESON2 316 is shown. This processing routine is started when a message is received on network 818 (determination Brock 316 of Drawing 20). Generally, this means that the message was received from re-transmission station 20. CRC error handling is very easy. A CRC error is recorded, and a message is thrown away completely and disregarded (Brock 344,346). When the thrown-away message is a message from re-transmission station 20, the following message which enters serves as sequence deviation (tested by determination Brock 348). A sequence deviation message is made at recorder station 28 where a demand is specified, in order to be temporarily stored in - memory buffer and to search the lost message (Brock 362-368). The lost message is pursued with message control Brock's 1354-1358 lost linked list to which it is pointed out by the pointer to lost message control Brock Liszt 1340. This lost message control Brock is shown in Drawing 59 in detail. When errorless, a message is inspected about whether it should be recorded or not (Brock 352). Only meeting 10,000 is recorded in this example (Brock 356-360). Drawing 23 -- processing routine MESONI -- 320 are shown. This processing routine is started by playback recorder station 26 when a message is received on network A16 (determination Brock 318 of Drawing 20). Generally, this means whether the received message is a thing from general participating station 24, and a thing from recorder station 28 specified. CRC error handling is very easy. A CRC error is recorded, and a message is thrown away completely and disregarded (Brock 384,386). When a message is a request message from general participating station 24, this message is disregarded, and it is processing routine ACKTM2. A response timer serves as the due date and general participating station 24 resends this request message so that 480 may be examined. In being a response message from recorder station 28 where this message is specified, this response message is disregarded, and in processing routine ACKTMO312, a response timer becomes the due date so that may be examined. Then, playback recorder station 26 resends this request message. the case where it is a thing from recorder station 28 where a message is specified -- (-- recorder station 28 where this is specified -- claim Was done -- it means supplying the right message information --) This right message is checked about whether it should be recorded based on meeting number field 52 (Brock 394). Only meeting 10,000 is recorded in this example (Brock 398.402). When playback recorder station 26 has still lost the message, (Brock 404.412) and this can send another request message for recorder station 28 (Brock 414) specified. When that is not right, the next request message for the completion flag 1346 of demand preparation and recorder station 28 specified is set in order to show whether it is behind, without waiting (Brock 420). When a message is a thing from general participating station 24 (tested by determination Brock 7390), (General participating station 24 loses a message, then this is from playback recorder station 26.) this lost message -- Doing -- via network A, playback recorder station 26 which means things sends a response message, and provides the right information (Brock (when available) 406-440). This demand is disregarded when a message is not available (in the Doing case [ For example, recorder station 28 which loses the message with same playback recorder station 26, then is specified this ]) (determination Brock 406). Since general participating station 24 is expected to require again after predetermined time, this can accept it, so that it may explain in processing routine ACKTM2480. Drawing 24 shows fatal error handling routine 298 about playback recorder station 26. With the fatal error which should be considered here, I hear that a response could not be obtained from recorder station 28 specified after the maximum number of times of a retry. As for me, saying [ a message being displayed and withdrawing a program ] hears that the whole meeting stops (Brock 428,430). Explanatory view 53-57 of general participating station 24 shows the data structure used by general participating station 24. Drawings 53 and 54 show main data structures. Drawing 55-57 expresses the details of a data structure. These data structures are the points of a format and are similar to the thing about re-transmission station 20. Drawing 25-33 expresses the detailed operation of general participating station 24. The main functions of a general participating station hear the message sent to recorder station 28 on network 818 specified from :(1) re-transmission station 20 as follows, and determine from it whether the message should be passed by the response pro galley. In this example, each message which has meeting number field 52 equal to 10.000 is displayed. (2) Detect all the messages that were lost or were able to be distorted and obtain the message righter than playback recorder station 26. (3) Using meeting 10,000, send a message to all other general participating stations 24 (it sends). In this example, two or the general participating station beyond it is used. Drawing 25 shows the flow of the overall system logic about general participating station 24. If general participating station 24 is started, in order to obtain all the messages about former meeting 10,000 where general participating station 24 is started, a temporary processing routine will be performed from playback recorder station 26 (Brock 442). However, main processings are performed in dispatch of processing, or weight loop 444 portion. When a user does not want to end the session and there is no fatal error, dispatch of processing or weight loop 444 portion is repeated repeatedly. A program is withdrawn, when it thinks that a user wants to withdraw or (Brock 446) a fatal error occurs (Brock 450). The effective message by which the last reception was carried out is stored before a program withdraws (Brock 454). Drawing 26 shows initial-setting 442 portion of processing in detail. The 1st initial-setting process comprises the following. - Message control Brock for RTC on network 81434, Message control Brock Liszt who comprises message control Brock for RTC on network aluminum 1436, ..., message control Brock for RTC on network An 1438, and message control Brock for PRC1440 is set up; -; which sets all the counters in a data structure as zero -- set up the completion flag 1504 of - transmitting preparation truly (this flag is used in order to control the flow of communication between the general participating station on network A16, and a re-transmission station). Other flags of all the set all the pointers other than the pointer to; and message control Brock Liszt 1432 set as an imitation as zero. After a station evacuates a meeting for a certain reason, a data structure is periodically recorded on a disk so that it can search at any time, when joining a meeting again. For example, general participating station 24 obtains first the message which was not received for meeting 10.000 from playback recorder station 26 at the time of a standup (Brock 462). These messages are displayed from it (Brock 464). The input from a keyboard or other input devices becomes usable after that (Brock 468). Drawing 27 shows dispatch of processing, or weight loop 444 portion in detail. The event which should be processed by the general participating station maintains the response timer about re-transmission station 20 on :(1) network B as follows, and measures the time of this becoming the due date. After re-transmission station 20 sends a message to re-transmission station 20 for broadcast of general participating station 24, I hear that that a response timer becomes the due date could not return a response message. The processing routine which processes response timer processing is called ACKTMI476. (2) The further response timer about playback recorder station 26 on network A16 is maintained, and the time of this becoming the due date is determined. As for saying [ that this timer becomes the due date ], after general participating station 24 sends a request message to playback recorder station 26, I hear that playback recorder station 26 could not send a response message. This processing routine that processes this further response timer processing is called ACKT M2480. (3) A message is received on network 818. Generally, this means that the message was received from re-transmission station 20. A message is checked about whether this has an error by the routine called MSGON2484. When errorless, meeting number field 52 is inspected about whether a message should be displayed or not. In this example, only the message belonging to meeting 10.000 is displayed. (4) A message is received on network A16. Generally as for this,; or the (b) response message into which :(a) response message which means the following things went from re-transmission station 20 entered from playback recorder station 26. The processing routine which processes these messages is called MSGONl 488. (5) The input from a keyboard or other local input devices must be processed. ; or the (b) user who decided that :(a) user who means the following things generally ended the program typed this in the message which should be sent to all the general e Addition stations 24 via meeting 10.000. The routine called KEYINP 492 processes user human power. Drawing 28 -- processing routine ACKTMI -- 476 are shown. This processing routine is started when the response timer about re-transmission station 20 becomes the due date (determination Brock 474, Drawing 27). This means not having been received, after a response message sends the message for broadcast of general participating station 24. It is resent unless a retry count is exceeded to the utmost (determination Brock 498) (Brock 502). Exceeding a retry count to the utmost means that it cannot communicate with re-transmission station 20 because of the object of transmitting a broadcast message. In this example, it considers that that error is fatal (Brock 500). Drawing 29 -- processing routine ACKTM2 480 is shown. This processing routine is started when the response timer about playback recorder station 26 becomes the due date (tested by determination Brock 478 of Drawing 27). This means that a response was not received from playback recorder station 26. It is resent unless a play pack recorder station retry count is exceeded to the utmost (determination Brock 510) (Brock 514). Exceeding this count is the object of recovering the lost message, and I hear that it cannot communicate with a playback recorder. In this example, it considers that that error is fatal (Brock 512). Drawing 30 -- processing routine MSGON2 484 is shown. This processing routine is started when a message is received on network B18 (tested by determination Brock 478 of Drawing 27). This generally means that the message was received from re-transmission station 20. CRC Creation-processing is very easy. A CRC error is recorded, and a message is thrown away completely and disregarded (Brock 524*526). When a message is a thing from re-transmission station 20 and is disregarded, the following message which enters serves as sequence deviation (determination Brock 528). A sequence deviation message is temporarily stored in a memory buffer, and in order that a request message may search the lost message (Brock 540-546), it is sent to playback recorder station 26. When errorless, a message is inspected about whether this should be displayed (determination Brock 532). Only the message for meeting 10*000 is displayed in this example. Drawing 31 -- processing routine MSGONI -- 488 are shown. This processing routine is started by general participating station 26 when there is a message received on network A16 (tested by determination Brock 486 of Drawing 27). This means that the response message was received from re-transmission station 20, or the response message was received from playback recorder station 26. CRC error handling is very easy. A CRC error is recorded, and a message is thrown away completely and disregarded (Brock 554-556). When a message is disregarded by the response message from re-transmission station 20, a response timer becomes the due date and m general participating station 24 will resend a message so that it may explain in processing routine 476. When a message is a response message from playback recorder station 26 (tested by determination pro Poke 558) and this message is disregarded, it is processing routine ACKTM2. The response timer explained in 480 becomes the due date. A request message is resent by general participating station 24. When a message is a response message from re-transmission station 20, it is (determination Brock 560), (It explains in KEYINP 492 like) The following perfect message typed from the keyboard is sent to a re-transmission station for broadcast (Brock 582). When a message is a response message from playback recorder station 26, (it being tested by determination Brock 558) and a message are checked about whether this is a thing for meeting 10.000 (Brock 564). A message is displayed when that is right (Brock 566). Drawing 32 -- processing routine KEYINP -- 492 are shown. This processing routine is started when there is a completion message inputted from the keyboard (tested by determination Brock 490 of Drawing 27). A completion message is a message which should be broadcast using the conclusion (Brock 600.602) of a program index, or meeting 10,000, and it obtains it (Brock 604-612). Drawing 33 shows fatal error handling routine 450 about general participating station 24. The fatal error which should be considered here is being unable to obtain a response message from re-transmission station 20, or being unable to obtain a response message from playback recorder station 26 after the maximum retry. Fatal error handling is very easy. A message is displayed and a program is withdrawn (Brock 618). In this case, note that the next processing step 452 of Drawing 25 stores the last effective message. This stored information achieves the function as an initial value, when a program is started again. Another desirable example of the present invention is shown in Drawing 34. This 2nd example differs from the 1st example in that it is said that this includes management station 22. The function of this management station 22 processes the start of :(1) meeting which may also contain the following, and a stop. Meeting number 1 is specific in this example. This is sometimes called "a global meeting", and in order to tell other stations of all the about information, it is used by management station 22. Such information may also contain the following: (a) Require all the playback stations to identify yourselves, those characteristics, and capability (memory capability, meeting number which these record, etc.). (b) Tell the start of a new meeting. And it is included in an announcement message which playback recorder station 26 records a meeting. [ being invited as the writer and leader of a meeting number, meeting start sound information, the name of a meeting, easy explanation of a meeting, and a meeting ] (c) Tell the conclusion of the present meeting. A meeting can be ended from one of meeting administrators with a clear demand to a management station, the due date at the time of a meeting end date, and the command from a management station; (d) Tell the event which is not expected. Use of the whole network must be made impossible for the problem in the reason for newest-izing of software and hardware, or communication or an electric field. (2) General participating station 24 which is going to start approval of a new meeting and a new meeting must obtain a meeting number from management station 22, in order to avoid inconsistency of a meeting number. (3) Processing of a secret conference. The thing with a meeting may hold secret information, and only the user who was able to lodge power can write the information, or can read it. A step required in order to process a secret conference is as follows. : (a) In the desirable example, from meeting number 10 to 9*999 is reserved for the secret conference. Re-transmission station 20 enters from management station 22, and accepts the message for these meetings only in a Came case. I hear that the message in which saying has a meeting number of one range of 10 to 9,999 of general participating stations 24 is refused. (b) One of general participating stations 24 must be beforehand registered into management station 22, in order to transmit or to be able to receive the message about a secret conference. This means that these must provide management station 22 with the information on front Around specification. Such information is not sent via a network from the reason of a safe side. Such information may also include other information used for the encryption / decryption key for communication with a node, informer discernment, an authorization level, and management station 22, and the object of attesting a user. (c) General participating station 24 (1) is general participating station 24. (5) to 24 It is assumed that it is going to make the secret conference rise with (7). This communicates with management station 22 to the 1st. Management station 22 is general participating station 24. An attestation procedure is performed in order to ensure that (1) is right and it is not deceiving. Management station 22 is general participating station 24, in order to acquire the information how the name of a meeting, Liszt, the easy write-in person explained and approved, Liszt of the leader approved, and a meeting end from it. It communicates with (1). Management station 22 assigns the meeting number between 10 and 9,999 from it to this meeting. Management station 22 and general participating station 24 The data in the communication of 1 to 1 between (1) is enciphered using the cryptographic key registered a priori for positive communication. (d) Management station 22 sends a meeting standup message on global meeting number 1 from it, and tells the start of this new meeting (for example, meeting number 99). As for the information about the name of a meeting, Liszt of the easy write-in hand explained and approved, Liszt of the leader approved, and Liszt of playback recorder station 26, meeting A stand"; is contained in a Ri message. (e) When general participating station 24(5)-24 (7) hears that these are the approved leaders on this new meeting 99, these communicate with management station 22 respectively. - This attests them using an attestation procedure. Meeting 99 decryption key is supplied to general participating station 24(5)-24 (7). This decryption key must be used in order to decode all the messages about meeting 99. The same decryption key is also given to all the leaders about meeting 99. (f) General participating station 24 (7) is the approved write-in hand on this new meeting 99, and it assumes that this tends to write in this meeting 99. This sends the message enciphered by the individual cryptographic key beforehand registered into management station 22 before to management station 22. Management station 22 decodes a message by an individual key, and is a meeting about this! It re-enciphers by 99 cryptographic keys. This enciphered message is sent to re-transmission station 20 from it for the broadcast on network B18. Please write in about meeting 99 and a hand should be careful of not needing access to the decryption key for meeting 99. Although the leader of a meeting, attestation of a writer, a leader, and writer Liszt are not separated and approved by such scheme, prevention of the intervention in a meeting is attained by it. Actual encryption/decoding scheme attested and used will become peculiar to a company or application. (4) An error and/, or processing of statistical information. A non-fatal error or statistical information can be sent via a reserved meeting like meeting number 2. If management station 22 does so, it can analyze or display such information. In re-transmission station 20. specification, the example of other error handling is detection of failure of recorder station 28 carried out and playback recorder station 26. In particular, failure of a playback recorder station is detectable by sending a message to this periodically via the art of 1 to 1. Above, the functions which described management station 22 are some only examples of Half. Public [ for specific use and/or data encryption, and decoding ] and the changed type of Such as assigning and the above-mentioned art is also a priori possible in all meeting numbers because of use of an individual key scheme. It is also possible to supply some of functions of the above-mentioned statement or all to other stations. The additional function like network safe protection, alarm, and resource assignment is possible similarly. The art in which individual and, as for public key art, a key serves as a pair is referred to. When an individual key is used in order to encipher a message, a public key must be used in order to decode the enciphered message, or is the contrary. An individual key cannot be made to derive though the public key and encryption/decoding scheme are known. Single bus network C composition is used for the example shown in Drawing 34 instead of the double bus arrangement shown in Drawing 1 (of course, like the structure of Drawing 1, if management station 22 is easily incorporable also into a double thing, there is). Cost can be reduced by using a single bus network. However, in the broadcast communication to recorder station 28 specified from re-transmission station 20, it is necessary to compete with other 1 to 1 communications on the same network. Since it comes out, a possibility of making a mistake in a message increases as the quantity of traffic increases. in a certain network called Ethernet -- " -- there is a possibility of falling into the situation known as miserable jamming, slightly. This situation occurs, when a message is lost for the numerousness of the quantity of traffic. Recovery art introduces more network traffics, in order to recover the lost message. This additional traffic causes many lost messages more to it. For this vicious circle, it may become impossible using the whole network. A broadcast alarm message for the possibility of miserable jamming to reduce a glitzy and a recovery mechanism, The advanced flow control scheme between global replay of the message broadcast a priori and/or re-transmission station 20, and recorder station 28 specified, Or using a scheme of other kinds, it can lower by various methods by reducing loading to each network Segcent. The 3rd example The example of the present invention desirable further now is shown in Drawing 36. Re-transmission station 20 and station 32 which performs the function of management station 22 are included in this further example. The advantages of this arrangement are that cost is reduced and that the number of the active stations on a network is reduced. As for me, a disadvantageous point hears that the complexity introduced into this joint station 32 is added. Many of functions of management station 20 are selectable, and over many hours, it can add or can change them. Although these systems may be able to be equal to the loss of short management station 22, they become fatal [ the loss of re-transmission station 20 ]. If joint station 32 breaks down, a system will suspend an operation. It is possible to have two or a management station beyond it. For example, one is used in order to process a secret conference, and another side can be used in order to process all the functions. In using joint station 32, such a possibility disappears. The 4th example Another example of the present invention desirable now is shown in Drawing 37. Single station 34 which achieves the function of the both sides of playback recorder station 26 and recorder station 28 specified is included in this 4th example. The advantages of this example are cost's further being reduced and being able to reduce the number of the active stations on a network. The following are contained in the disadvantageous point of this example: (1) If the number of general participating stations 24 increases even to thousands, these tend to communicate with single station 34, in order that all may recover a message. A problem becomes serious at a Caught case about the message by which many such general participating stations 24 were altogether lost at the same time at the time of the start on the 1st. (2) As a separated station, some of the playback recorder stations 26 can also play a role of a file server to some of the general participating stations 24. This is impossible about joint station 34 in large scale network environment. However, station 34 can also achieve a file server function. The 5th example Another example of the present invention desirable now is shown in Drawing 38. What was combined with re-transmission and a management station by this example, and general participating station 24 (1) is combined and the thing used as joint station 36 is contained. Another general participating station 24 (2) combines with a playback recorder and the thing combined with the station specified, and it becomes joint station 40. To the advantage of this example, the number of the active stations on cost and a network may be able to be reduced. The complexity about :(1) joint station 36 where the following are contained in a disadvantageous point, and joint station 40 increases; (2) When one of the joint stations of these breaks down, it is that multitasking capability is needed about; and (3) joint stations 36 and 40 which the function of a system stops. The 6th example Another example of the present invention desirable now is shown in Drawing 39. In this example, the function of re-transmission, management, a playback recorder, and the recorder specified is combined with a single station. As for the advantage of this example, :(1) cost which is as follows does not need a response [ station / recorder / by which;(2); (3) specification which becomes low, and by which the number of active stations is reduced is carried out / 28 ]. Before sending out joint station 23 on network C30, it only stores a broadcast message. single Pointing; of the failure in;(2) joint station 23 in which whose complexity about (1) joint station 23 the following things are contained in the disadvantageous point of the example of this reduction composition, and increases (3) -- especially When many general participating stations 24 desire a message altogether lost when [ like the time of the start on the 1st ] the same, Since there is no response message sent from recorder station 28 where; (4) specification of the loading to a single station which becomes high is carried out, a certain transmission error is not detected easily. (5) The general participating station which is late or has a defect may damage joint station 23 by requiring the always lost message. The 7th example The 7th example has been especially changed so that satellite transmission may be suited. In this example shown in Drawing 77, it of recorder station 28 specified as the function of re-transmission station 20 is combined. The application of this example in satellite environment is shown in Drawing 78. Joint station 2000 is carried on a satellite. Playback recorder station 26 and general participating station 24 are summarized at a conspicuous place. Binding of each place is carried out to a satellite or the ground station which can transmit and receive information by this. Each place can record the meeting message similarly used by general participating station 24 at the place. It has at least one playback recorder station 26. This playback recorder station 26 is connected to the remainder of the nodes of that place via a local area network like Ethernet. In this example, general participating station 24 to which :(1) message whose step in connection with attaining the guaranteed reliable broadcast is as follows is sent transmits that message to a ground station. (2) A ground station sends the message to joint station 2000 on a satellite. (3) Joint station 2000 attaches special message header 44, and stores a raised message. Then, the raised message is broadcast. Joint station 2000 should be careful of not expecting the response message rather than which node in a network. It can be said that this broadcast message includes an error. (4) The broadcast message is received by the ground station and this broadcasts a message on a local area network after that. (5) Playback recorder station 26 inspects the broadcast message, in order to detect omission or distortion. When errorless, a message is thrown away or is stored according to meeting number field 52. It is To to joint station 2000 so that playback recorder station 26 may be re-broadcast via a ground station, when there is an error. (6) Joint computer 2000 does not form the 1 to 1 communication used as playback recorder station 26 and a pair, in order to supply omission / distortion message. This sends a claim Was done message at a global meeting (for example, meeting number 3). This is performed because many playback recorder stations 26 and general participating stations 24 may have dropped out the same message simultaneously especially when an error occurs at the time of the end of transmission of joint station 2000. If two more or playback recorder station 24 beyond it makes Caught the message same in the time interval defined as beforehand [ fixed ], Since an omission message is sent only once in this time interval, all the demand stations will have an omission message. (7) Playback recorder station 26 holds meeting 3 about the It was witnessed message. When the message demanded in the fixed time interval does not appear, this requires a message again. (8) General participating station 24 hears a message which is explained in Step 4, which is sent by the ground station and which was raised. When message omission or distortion is detected, this can hear global meeting number 3 from playback recorder station 26 first alternatively, before requiring the message which dropped out and which was able to be /Distorted. Thus, all the key elements of the present invention are saved in this example. Jam: (1) Detection of the message error for [ reliable ] attaining one pair of data communication of opposite a large number in large numbers or in large numbers, and recovery [ this ]; (2) They are a store and search of this about a message to playback recorder station 26 so that general participating station 24 may leave a meeting to To lose and can re-join it at all in data; (3) Ordering of the global message which enables in-series-ization of a database demand. : (1) whose advantage of this example is as follows -- this is fit for communication [ like / satellite-oriented ] especially. The delay relevant to a message response is removed. (2) In order to take into consideration the fact of being serious, delay of the transmission under such environment has many remarkable messages, and they obtain it. (3) By broadcasting an omission message, the possibility of miserable jamming becomes low. Since a good message must be temporarily stored while demanding, in order that the disadvantageous point of this example may recover; and (2) poor message whose :(1) complexity as follows increases, a large-scale memory buffer is needed. The 8th example This example is typically applicable only to a local area network. In this example shown in Drawing 79-81, most functions of re-transmission station 20 are included in raised recorder station 2100 which is specified. All the nodes have turned to this example only to broadcast or the local area network which can carry out a multiplex broadcast. This cannot be used in a satellite system. In this example, :(1) message-transmission general participating station 24 whose step in connection with achievement of the broadcast which can set Faith is as follows attaches message header 44 special to that message. ; (2) This broadcasts directly a message raised from it on a network. Then, it waits for a response message from raised recorder station 2100 which is specified. This raised recorder station 2100 that is specified has responsibility in maintenance of a global message sequence number, newest-izing, and offer of time information. (3) The response message sent from raised recorder station 2100 which is specified contains a newest-ized global message sequence number and time stamp information as data (Drawing 80). Special message header 44 is the same as the thing of the informer of a message. This response message is similarly broadcast directly on a network. (4) Playback recorder station 26 hears the both sides of a message and a message response. In order to measure whether the message has dropped out, this looks at the global sequence number first contained [ 1st ] in a response message. And a comparison function is performed in order to find the message corresponding to a response. When the response message has dropped out, (sequence deviation) playback recorder station 26 is, It is required that the art of 1 to 1 which can set Faith should be used for raised recorder station 2100 which is specified, and the both sides of the response message which dropped out, and a related message should be seen off in it. When the data message has dropped out, playback [ (it is determined when a response message is received, and there is no agreeing message header) ] recorder station 26 can decide whether to desire a message rather than the meeting number field. If it is necessity after that, an omission message will be required from raised recorder station 2100 which is specified. (5) This receives general participating station (computer) 24 by the same method as playback recorder station 26 except for the point of requiring an omission message from playback recorder station 26, and it detects a message error. Thus, all the key Elamation of the present invention are saved in this example. Jam: (1) The detection of a message error for [ reliable ] attaining one pair of data communication of opposite a large number in large numbers or in large numbers, and recovery [ this ]; (2) They are the store of the message on braver Poke recorder station 26, and search so that a meeting may be left and it can re-join, without general participating station 24 losing any data; (3) Ordering of the global message which enables in-series-ization of a database demand. :broadcast message whose advantage of this example is as follows is sent only once from the source of a message. Since broadcast of the node of all the :so as follows must be able to be carried out directly, it cannot use the disadvantageous point of this example for a satellite type system. : (1) whose disadvantageous point of this example is as follows --; (2) which all the node cannot use in a satellite type system since broadcast must be able to be carried out directly -- since all the node hears it about a broadcast message and the both sides of a broadcast response, complexity increases. Agreement between a broadcast message and a broadcast response message is needed. (3) Since many broadcast messages are before receiving a response message (it is one from each possible sauce), a large-scale buffer is needed. (4) Lack of a flow control mechanism. Raised recorder station 2100 which is specified cannot drop easily the rate of the message from general participating station 24 to which the message which a large number became independent of is sent. Another combination of various kinds of stations is possible. Probably, the example of these is a changed type of the example of Drawing 34 where playback recorder station 26 and the general participating station were incorporated as a joint station. Error recovery It is assumed that it is that for which the same hardware is used in network A16 of Drawing 1, and network 818. When network A16 breaks down, it can be used in order that B18 may perform all the network traffics. When network B18 breaks down, it can be used in order that A16 may perform all the network traffics. General participating station 24 may break down any time. A backup copy can be used in order to recover. A backup copy determines which is the message received correctly at the end first. And it is required that all omission messages should be sent to playback recorder station 26. Playback recorder station 26 operates all the time, and each can record one or the meeting beyond it. It is desirable for one or more playback recorder stations 26 to record the meeting of thing [ any ] specification. It does not become fatal [ an intermediary without one playback recorder station 26 ]. Playback recorder station 26 is recoverable using the backup for requiring that all the messages that dropped out to another playback recorder station should be sent. About re-transmission station 20, management station 22, and recorder station 28 specified, it is in former, Since a redundant hot standby station recommends, an electric signal must be sent to a satellite and time delay must be returned in satellite communication, it is serious. Network capability will become very late, supposing all the messages must be answered before the following message is sent. : which recommends the following improvement measures -- the group of a message can be sent before receiving a response from station 28 where (1) re-transmission station 20 is specified. This is known for the data communication industry as '' communication window" art. (2) Put re-transmission station 20 on a satellite. Thereby, transmission delay can be decreased. One problem in the case of putting a data processing station on a satellite is exposed to a cosmic ray (the atmosphere of the earth absorbs most of such light). It is expected that an error ratio becomes higher. (3) Recorder station 28 specified can be similarly put on the satellite near the same satellite-izing or this. A response message can be sent soon farther. An alternative plan is performing re-transmission and the recorder function specified at the same station. (4) Each station does not need to have satellite communications skills. A station is summarized and can be connected using low power and an area network. One of the stations in a local area network has satellite communications skills, and it processes meeting traffic for other stations in the same local area network. For example, in a token ring network as shown in Drawing 35, before it is considered that transmission of a message is completion, the message sent out by the station of either of the rings will carry out round travel of the ring completely, and will return to the original informer's place. If each station in the same token ring network provides the hardware changed slightly or a microcode even when a message does not make that a destination, it can hear the message. If it becomes receivable [ the message which is not specified about the station ] by a microcode (that which is already usable to network management nodes), Sending can be supported efficiently at every station. Use re-transmission station 20 of the hardware for improving the efficiency of a system does not need the 2nd memory storage. It is appropriate to carry out a "black box '' type for exclusive use. Recorder station 28 specified does not need to need the 2nd memory storage, when it has sufficient memory buffer space. A message accepts it temporarily and is stored so that sufficient chance for playback recorder station 26 to recover the message which dropped out may be welcomed. Implementation for exclusive use "black box ' type can be taken into consideration. The communication interface on network 818 may need a lot of message filter rings. As for saying, I hear that these can permit only the message which has a specific meeting, a specific informer, or the specific message field. It is the best to perform this type of filter ring by the hardware which uses the present art. The concept of the guaranteed reliance The concept of the guaranteed reliability is important. In data communication, since a possibility that a message is lost or it can distort will become very low if displayed on an application program, that the message was transmitted well is like [ which can disregard this for all the objects on practice ]. For example, the telephone line is usually said to be a bit error ratio of 10-4. When the average size of the message which should be transmitted is 103 A bit, the possibility of distortion of a message will tell 1 to 10. This problem is conquered by use of a communication protocol. In order to detect the loss of a message, use a message sequence number. (1 has changed to detection of the message which was able to be distorted at O, or the reverse) and the art known as a Cyclic Redundancy Check (CRC) are usually used for a target. He can understand CRC art by comparing this with the concept of arithmetic division. Is, a quotient, and remainder come out by division. - In data communication, the data flow which should be sent out is a number which should be broken (the number of Removal-ed). The number with which the number of Removal is chosen specifically (known as a CRC polynomial). after performing division, the number of Removal is disregarded -- too much -- " -- equal number ' is added to actual data flow (known as a CRC numerical value). At a receiving side, it is the same division mosquito line Wow (the same CRC polynomial) by the same number of Removal after receiving data flow. When errorless, remainder must be the same as a CRC numerical value. By choosing the size and the numerical value of a CRC polynomial properly, the probability of receiving the message which was able to be distorted can be kept very low. When only 1 A bit has made the error, it is clear that the error is detectable. About a multiplex bit error, there is a possibility that a certain error will not be detected, slightly. However, the possibility in the 32 A bit CRC art at the time of being applied to the example of the telephone line of the above-mentioned statement will call it one message which is not detected in about 16. Thus, the possibility of a such error is disregarded about the practical object. The message is resent whenever an error is detected using CRC art. The noise of a communication line is too high, and it is stopped by message transfer when an exact message is not received after the retry over several times. An application program is told that a message could not be transmitted. Thus, a word called a guarantee is 100% of reliability at 1 real target which means that transfer of the message for all the common intentions or the objects goes well, when used by data communication. In the context of the present invention, a word called a guarantee has a similar meaning. However, all the playback recorders may destroy destructively all the data recorded by causing failure. The meeting participating station which is using it or turned off the switch may not obtain a broadcast message. This possibility is considered to be such a low thing in practice that it can ignore. A playback recorder can be added when higher reliability is required. Flow control One important concept in data communication of 1 to 1 is a flow control. This is art for a message transmission data processing station to reduce the rate which sends a message (or it accelerates). Since it cannot follow but the loss of a message may produce a receiving data processing station when a message transmission data processing station sends out a message early too much, this is important. A tacit flow control mechanism is provided by message response in the data communication environment of 1 to 1. When a receiving data processing station is late or in use, this only delays transmission of a message response. In the case of the present invention, re-transmission station 20 is to many playback recorder stations 26, When sending out a message early too much, these playback recorder stations 26 form the communication link of 1 to 1 with recorder 28 specified, in order to require an omission message. This information can be used for recorder 28 specified in order to change that message response delay timer. When many playback recorders require an omission message saying, Recorder station 28 specified reduces the flow of the message from re-transmission station 20 by increasing the time interval between reception of the message from re-transmission station 20, and transmission of a response message. A flow control is attained by this art. In addition, some general participating stations are in use, or when late, an omission message comes to hand from one of many playback recorder stations 26. When playback recorder station 26 can process such a demand, the flow of a message is not affected at all. This solves one of the main worries in the mixed network environment to which the data processing station of an extensively different speed is connected. The informer does not need to have the latest recipient. Operation strategy In carrying out the present invention, there are many alternative methods. (a) All the example of application is dealt with independently. About the special example of application, a program is written specifically. Since it is the easiest to understand this approach and the easiest to carry it out, it is taken up by the 1st example. (b) Write the set of the common subroutine for application programmers. a system programmer defines an application program interface -- the basis of an application programmer to all the protocols and the complexity of network composition -- " -- ' to hide. An application programmer publishes the subroutine call of using this application program interface. These subroutines are linked to his program image. This approach is taken under the examination about use of the present application. This approach makes it very easy that an application programmer programs. They do not have even the necessity of knowing how the present invention acting. (C) Incorporate this invention in an operation system as "communication surveillance." This is the most effective method as far as use of a memory and the efficiency of a code are concerned. However, it means that this changes the operation system developed by different company. This approach should not be taken as long as there is no very intimate cooperative relation with the company which develops / maintains an operation system in Okay.. Coexistence with the conventional system The present invention can be coexisted with the conventional system via use of "translator node '. For example, an electronic mail message shall be sent to another node of network 1,000. Other things can use only the conventional art of 1 to 1 to the ability of a certain thing of the nodes of these 1.000 to use the present invention. The informer can use the conventional mailing list distribution art except for the "translator node" being contained in a mailing list. A "translator node" uses the present invention after receiving a mail message based on conventional 1 to 1, and broadcasts this. In this way, coexistence is attained. the subroutine for application programmers -- in the detailed step examined in the desirable example, a system programmer is required for actual operation. If there is one of the usual art of this industry (the system programmer does not need to conduct the experiment of : and an unnecessary quantity, uses the example programmed control step shown in Drawing 5-33, and the data structure shown in Drawing 40-59, and could set a subroutine easily.) The application programmer does not need to be worried about how the present invention operates, and can use the set of this subroutine. The following is an example of the set of such a subroutine. (1) SETUP-CONFERENCE (PI, P2 *P3 ... P9) It is here at the due date time of a type (is secret? isn't that right?) P2-meeting of Pl-meeting. The pointer to meeting number numerical value P5-controller Liszt who returned in the P3-success or the failure indicator P4-success (for example, what can carry out a meeting stop, the thing which can change Liszt of a controller, the permitted writer, or a leader) The pointer to Liszt of the leader to which the permission was granted when the pointer P7-meeting to Liszt of a writer to whom the permission was granted when a P6-meeting was secret was secret - The pointer to the alternative message broadcast with a P8-meeting start message - Pointer (2) STOP-CONFERENCE (P2.P here PI, Pl-meeting number) to the name of a P9■ meeting The pointer to the alternative message broadcast with a P2-meeting stop message - A P3-success or failure indicator (3) WRITE-CONFERE NCE (PL, P2 *P3) It is PI-meeting number here. P2 - Point Tarr P3-success in the message which should be broadcast, Or failure indicator (4) READ-CONFEREN CE (PI) P2.P -- here -- Pl-meeting number P2 - a pointer P3-success to the buffer in which the received message should be stored, or failure indicator (5) SET-MASK -- a pointer [ here (PI, P2.P3) as opposed to a Pl-meeting number P2-message mask ]. A mask may be based on an informer's name, message type, or message content value. a P3-success or failure indicator (6) MODIFY-CONFER ENCE -- here (PI, P2 ... P5) -- Pl-meeting number P2 - Liszt of the Liszt C meeting leader of the Liszt b meeting writer of the type , More a meeting controller of the information which should be changed A pointer P5-success to Liszt specified as a P3-addition or deletion indicator P4-P2, or failure indicator (7) READ-G LOBAL-EVENTS(PI, P2.P3) PI - The number of unsettled global events (message) the [ P2-] -- in the use outline portion of a pointer P3-success to the buffer which stores the unsettled global event (message) of one, or the failure indicator 1 present invention It is shown how about various kinds of examples examined first, especially I will reinforce here, using the program interface of the above-mentioned examination, these examples are carried out by the application programmer. 1, a stock market trading system In the conventional system, the information on (Drawing 60 Reference) and various kinds of stock markets was inputted into the single data processing station of the securities firm. The customer who asks about a stock price or places an order for purchase and sale accesses this single data processing station. This single instruction single data stream station cannot respond to a busy day at the peak hour of an order in particular. It was not able to receive, even when a lot of dealings were predicted several days ago. Thus, since there is a serious fault in the conventional system, it is not easy to extend this and to accept a new user. In the case of the present invention, the information on various kinds of exchanges is broadcast using the data communication art of guaranteed reliable one-pair a large number. By one's personal computer or workstation, the representative of a broker or a customer can obtain the information broadcast alternatively, and can display and analyze it. When a lot of dealings are predicted, more staffs and a personal computer are added and a demand of visitors, such as an inquiry of a stock price and an order of selling and buying, can be coped with. This concept should set the following by an application programmer at the central place where :(a) stock price which shows how it is feasible is broadcast. Varnish Step 1 whose main processing steps in connection with broadcast of a stock price are as follows Stock quotations are obtained; Step 2 CALL WRITE-CONFERENCE broadcasts a stock price, and it returns to Step 1 after that. Cb) In a broker or a customer representation workstation, the main processing steps in connection with reception of the stocks broadcast set up a filter peculiar to a customer by varnish Step I CALL SET-MASK as follows; Step 2 CALL READ-CONFERENCE receives information; Step 3 A message is processed in a buffer and it returns to Step 2 after that. This can be compared with a meeting of the human being in whom one person's speaker is. It is very easy to increase the number of hearers, without affecting a speaker or a meeting. As far as a speaker is concerned, it is also the same to also talk to 300 hearers and to talk to 1,000 hearers. the present invention was lacking in the prior art -- expandable solution is provided easily. 2, software distribution With the conventional software distribution system, they are software and/, or rise boots, A network is passed and it is on the base (not saying) of 1 (instead of distributing diskette or same thing) to 1. reliable reception is absolutely [ software rise boots ] important for the proper function by the side of the preservation nature of a system, and a receipt -- it can send electronically. In the case of this art, there is a fault that the close cooperation between a software distribution data processing station and a receiving data processing station is needed. I hear that distributing the same software to the receiving station of 1,000 sends the same information individually 1000 times. For example, a certain company will assume that a new software application program is made into all the offices 1,000 in the world with distribution use. Supposing one transmission takes an average of 15 minutes, transmission to all offices will take ten days or more. - in which delay of this time may not be accepted is because the release (an old version and a newest-ized version) of different software may be unable to live in a network together. In the case of the present invention, software and/, or rise boots can set the guaranteed reliance, and it can send them using the positive a large number art versus 1. If wished, only the user who was able to lodge power can ensure being transmission and being/, or that it is receivable for rise boots with the secret conference characteristic which uses the management station of the present invention. A software distribution data processing station does not need the close cooperation with many receiving data processing stations. In a desirable example, I hear that distributing the same software to the receiving data processing station of 1,000 should send only twice under environment without an error of the same information. Required time will be not the 10th day or more but about only 30 minutes. Receiving perfect and exact information is guaranteed to the receiving data processing station. Hereafter, this concept shows how it can carry out by an application programmer (refer to Drawing 62). (a) At a software distribution place, main processing steps are Step 1. It reads in Brock with the software which should be distributed. Step 2 CALL WRITE-CONFERENCE broadcasts the Brock. Then, it returns to Step 1 until the whole software program is broadcast. Step 3 END OF FILE MESSAGE is broadcast. (b) At a receipt place, main processing steps are varnish Step 1. Step 2 which switches on a system (place; which is not [ of a system ] on yet) Step 2 is repeated until;END OF FILE MESSAGE which receives software is received by CALL READ-CONFERENCE. By 3, education, or the system of the demonstration former, a teacher shows students the output which appears in the screen of the terminal or a personal computer. In order to show more audiences the information, he may use a projector alternatively. In this arrangement, it is required for all the students to be in the same classroom or the same time zone. In the case of the present invention, the teacher can use the a large number art versus reliable 1 to which the output of the screen was guaranteed, and can send as a broadcast message. Students can display this screen output on their personal computer. They do not need to be in the same classroom. The student who "cannot participate in real time" in - and the lecture which do not need to be present even in the time zone where students are the same for the record capability of a playback recorder station can see this on next time and/, or a day. Since only a screen output is displayed, students do not need the demonstration software in the personal computer. This situation receives TV left camera top demonstration, and displays this on a plurality of television receivers simultaneously, and can compare it with storing in a VCR. Special demonstration software Or special demonstration hardware is because the fact which will be said if unnecessary at a student's receiving station can participate only by seeing a display screen also by serious -, because the student that has only a system of composition at worst. Probably a special advanced large-scale mainframe computer with software is used for a screen, and it could display it on a domestic personal computer by - and the students who are transmission sides and will be able to create. Thus, in the side which is that for which the present invention paves a demon strike and the new way which carries out a rate for software and which is seen one, The necessity for the hardware used in order to create demonstration graphics and a display and/, or software is made for there to be nothing, The screen output with which the demonstration was received, The screen which could display on students' station in which many carried out remoteness using communication of reliable one-pair a large number, and was displayed is voice, a picture, graphics, or a similar thing in advance of a display, and edit or reinforcement is possible for it. :(a) teacher side who shows how this concept may be hereafter carried out by the application programmer (refer to Drawing 63), ; which performs the following things --; which adds the following subroutine calls before sending a certain information to sign start 2 screen of the demonstration by DEMO5TARTMESSAGE with 1 CALL WRITE-CONFERENCE -- for example CALL WRITE-CONFERENCE, CALLWRI T E to5creen ; After reading from three keyboard, the following subroutine calls are added; CALL READ froa+keyboard and cALL WRI TE-CONFERENCE. 4) It is a sign about the end of a demonstration by DEMOEND ME SSAGE in CALL WRITE-CONFERENCE. b) D which performs the following things in the student side Step 1 It is made to synchronize with DEMO5TART with a study screen etc.; Step 2 Screen information CALL- and READ-CONFERENCE are obtained; Step 3 The received information is sent to a screen. To for example, CALL WRITE toscreen or CALL WRI T E to disk ; which stores Information -- until END OF DEMOMESSAGE is received Step 3 is repeated. In the system of the communication former with the receiver group of 4o strangeness, It only desired for a strange receiver and a data processing station to carry out group transmission to have an active receiver of the recipient who only the broadcast transmission which reliance cannot set in peculiar can use it, but is meant, and for the message to reception be ready for the message. In this system, reception of a message cannot be guaranteed and cannot set reliance essentially. In the case of the present invention, the informer can use the data communication art of guaranteed reliable one-pair a large number based on a meeting number, and can send the message to it. It is not necessary to know coincidence of a receiving data processing station. That a receiving data processing station hears a broadcast message in real time or acquiring exact information via a playback recorder data processing station are guaranteed. A receiving data processing station will ask whether it participates in use of the above-mentioned meeting number. The receiving data processing station can participate, also after a meeting starts and many days or many weeks pass. '' delay subscription "-- requiring a receiving station of the record capability of a playback recorder station -- this -- " -- lost "information is made to provide In :(a) meeting start which shows how this concept can carry out the following by an application programmer (refer to Drawing 64); It is meeting start D by step I CALL 5 ETUP-CONFERENCE. Transmission of the message by step 2 CALL WRITE-CONFERENCE; (b) With a receiving side; the standup of a meeting by step I CALL READ-GLOBAL-EVENTS -- detection; -- step 2 CALL READ-CONFERENCE -- data reading; It is the writing to a meeting by step 3 CALL WRITE-CONFERENCE. 5 Network management Each data processing station on a network was individually controlled by the conventional system. One disadvantageous point in this arrangement needs to change a global parameter, Or when the patch to all the software or a certain thing needs to be installed, I hear that the cane of the change must be carried out to each data processing station on a network, and it must be individually performed to it to it. By this pro, about a large scale network, A also costs every day or how many week -, and all network users' cooperation is needed for it. In the case of the present invention, it can change or patch at the time of Okay. said with the art of reliable one-pair a large number in which the whole network was guaranteed (and further). The secret conference capability provided by a controlling function/station if wished can be used, access to the change can be restricted to a specific station or a user, and it is 1. It is not necessary to worry about the network which has some of data processing stations which did not need to leave the possibility of omission, and did not need to change each data processing station, were changed, or were patched to work. Each receiving data processing station answers the reliable positive message to which the network software or a parameter was guaranteed, and it can change or patch it independently. It is necessary to change some of global parameters, and must perform the patch to software for this application. Changing an execution system directly is not usually put into Acceptance. It is more nearly usually obtaining the copy of new software on a target. If a system is booted again, this new copy will be taken and replaced with an old thing. the following -- this concept -- an application -- in the =(a) management side which shows how a cage gin programmer may carry out (refer to Drawing 65) --; step 1 -- the copy of this new software is broadcast as a new file. CALL WRITE-CONFERENCE. Step 2 In; (b) receiving side broadcast on the batch file which defines the step which installs or performs software CALL WRITE-CONFERENCE; Step 1 --; which receives copy CALLREAD-CONFEREN CE of new software --; which receives step 2 batch-file CALL READ-CONF ERENCE Step 3 A system is rebooted, the received batch file is performed, and the copy of new software is installed and performed. In order to carry out the trap of a user and the invader, it was required to change a standard operation system. This art will apply additional and another overhead requirements to a data processing station. A certain history data is made more difficult in recovery than one error which may not be recorded on a period busy for an overhead, accidental file elimination, and the same thing. In the conventional system, as for safe protection, what is provided as afterthought is frequent, and only after a safe safeguard also has an invader's doubt arisen, it is introduced. Many damages may have occurred by this time and there are not character of that damage and a method of judging a grade, either. In the case of the present invention, all the inputs to one of data processing stations (or all) and all the change in the database of a data processing station can be broadcast and recorded using the data communication art of guaranteed reliable one-pair a large number. The recorded message achieves the function for providing historical record like backup. Even analysis or replay is more possible for an invader's activity than this history record. Record makes mind of being a design stage of - system provided as an indivisible portion of an overall system function as a matter of fact, and making safe protection monitor capability containing cause. The safe protection analysis program can also analyze the event which occurred before having become behind, and being able to add thru/or change at any time and actually introducing these. The following shows how the concept of two may be carried out by the application programmer in the data processing station which should be monitored (refer to Drawing 66). It substitutes for a READ DATA subroutine by READ DATA and CALL WRIT E-CONFERENCE so that the input data of step 1 all so may be broadcast, stored and monitored; It substitutes for a WRITE To DISK subroutine by CALL WRIT E-CONFERENCE WRITE To DISK so that all the change to step 2 database may broadcast, store and monitor. 7 Computer conferencing In the conventional system, the computer conferencing between many data processing stations was attained by requiring all the meeting participants to read and write from a single database. In this arrangement, a Overpotential load will be applied to the data processing system which often maintains a single database. In the case of the present invention, the art of a large number versus guaranteed reliable a large number can be used. No meeting participants need to write in a single database. The meeting participant can transmit or receive a message, without worrying that an overload may be applied to a single database data processing station. 1 which can even perform a meeting participant standing on the message broadcast before [ a certain ], and returning -- for a new meeting member, it is the especially useful feature. The new member can search the former message sent about the meeting from a playback recorder station. The following shows how this concept may be carried out by the application programmer (refer to rI!J67). It is data transmission by step I CALL WRITE-CONFERENCE; It is data reception by step 2 CALL READ-CONFERENCE. 8, a support of a different company Especially the conventional network had it assumed to exist under friendly environment, when a network was based on broadcast or multicasting capability. In order to exchange restricted data under such environment, encryption of data and decoding are performed on the base of 1 to 1. - Many cooperation and an overhead are needed. In the case of the present invention, the alternative certainty of the art of a large number versus guaranteed reliable a large number and the restriction feature can be used, and access of specific communication can be restricted to a specific user. The company which supports will assume a special broadcast network to supply in order to support the company customer of the many. When the company which supports wants to advertize the product, this can be broadcast at the meeting that only the company which supports can write in, although all participants can read information. When a customer company wants to send information to other companies, the message to leave can carry out restriction or course specification via an alternative management station. before this management station broadcasts this -- a monitor and/-- or it can edit (for example, before language which is damaged, or a complaint broadcasts, it can be processed first.) A good expression can be broadcast to it, without taking prior action. The key for carrying out this application is use of a secret conference. An application programmer's use of a secret conference (for example, % 200) will carry out course specification of him automatically to broadcast Messe management station 22 of this. Since it comes out, an application programmer's task is very easy D (refer to Drawing 68). step I CALL WRITE-CONFERENCE -- message transmission (at meeting 200); -- step 2 CALL READ-CONFERENCE -- message reception (at meeting 200). 9 Backup system By the conventional system, in order to '' backup ' Carry out the information in which this stores a data processing station, it may be necessary to operate it in self-education mode (when it is got blocked and the 1st memory storage of a station breaks down, information is copied by the 2nd memory storage). In the arrangement of two, the front Around schedule of the backup needs to be carried out, and while a data processing station backs up, there is a fault that it needs to be operated in the independent mode. When the data processing station backed up is central to network operation and independent equipment is not arranged, the whole network may stop for a while. In the case of the present invention, using guaranteed reliable one-pair a large number or the art of a large number versus a large number, all the change to all the inputs to a data processing station and the database of a data processing station can perform broadcast and record. Information required in order to reproduce a backup copy is always usable in a playback recorder station in this way. The copy of a data processing station can carry out [ newest ]-izing without a prior schedule always, and the independent mode is unnecessary. Perfect backup of main data processing stations is performed in the independent mode at the time of :(a) predetermined [ whose a detailed step is as follows (refer to Drawing 69) ]. Each of copies, then prior arts which was said can use the whole contents of the disk for magnetic tape. The predetermined image of a data processing station is supplied by this backup copy at the time of predetermined. (b) Using guaranteed reliable one-pair a large number or the art of a large number versus a large number, all the change to all the inputs to main data processing stations and the database of main data processing stations is broadcast and recorded. In practice, the playback recorder station has already provided record and a memory function. (C) When main data processing stations break down, the image stored in magnetic tape is copied to another data processing station of hardware constitutions similar to main data processing stations. (d) The recovery process is almost the same as the rerun of all the processings performed by main data processing stations. The memorized input (or memorized meeting message) is searched from a playback recorder station, and the output to which a backup data processing station is turned to a terminal, a printer, etc. performs processing except being sent to an empty device here. It is only the memorized input messsage and a backup system can be theoretically returned to the same operation state of the main data processing stations in front of failure. There is a phenomenon known as "database drifting carrying out only (e). That is, the same input is received and at least the two same systems that run the same program have that after a certain period is not the same (for example, interference [ exterior / the slight difference in I\-Dwarf, the factor of an environmental side, and ] or bombing of a cosmic ray). In order to detect such "database drifting '', the memorized change message to the database of the main data processing stations which made reference at Step b is used. When a backup data processing station needs to make a change to a database, this always compares with the change message which had the change memorized. In this way, any database drifting is detectable. And human intervention is To be because of the analysis effect of a database drift. Human intervention is important. It is because possibly the defect was caused by main data processing stations (for example, a cosmic ray will bomb an electric circuit by chance, and will change a bit into 1 from zero, without doing eternal damage to hardware). In Okay., difference is noticed and the change to a database continues backup as what was performed by the main database processing station. This is analyzed later. varnish Step 1 by which actual execution is attained as follows -- the input data of all the so is broadcast and memorized -- as -- a READ DATA subroutine -- READ D ATA and CALL WRITE-CONFERENCE -- substitution. Step 2 It substitutes for a WRITE To DISK subroutine by CALL WRIT E-CONFERENCE WRITE To DISK so that all the change to a database or change may be broadcast and memorized. 10, a hot standby system In the system of the former for which a reliable operation is needed absolutely without down time, an excessive (main and the 2nd) data processing station is arranged typically. A main day evening processing station has responsibility in actual data processing operation. The 2nd data processing station (continuously, all the communications are monitored, therefore it is always the "present") is succeeded when main stations break down. Typically, and it is interlocked closely (for example, specialization hardware and software), and the 2nd data processing station is difficult to add the 3rd backup system. In the case of the present invention, a hot standby system can be designed by a method similar to the backup system of the above-mentioned examination. One which functions as a ' hot standby ' station, or the data processing station beyond it, Each "database drifting is detected as compared with the change message which could hear the same input, could perform processing of; resemblance, and was broadcast about compatibility in; and the change as a result. Since what is necessary is to do the work like a payroll payment check only once, it is similar, but the processing which is not the same is emphasized. A term called "shadow wing is used by the present invention. In the case of the present invention, any number of hot standby systems can be arranged simultaneously, and it can add and can delete them at any time. A hot standby system should be careful of being the backup system form examined above in practice. "The recovery process of the only difference" is [ one or more ] only carried out substantially in the same data processing station again in real time. Although the one-person Khote sage in which the failure of main stations cannot answer to :(a) input messsage examined in either of the following situations is defined a priori and number-of-times resending of the maximum is carried out, each - trial is performed after predetermined timeout duration. (b) A database change message cannot be sent after the input reception by active main data processing stations. If failure of main data processing stations is detected, it will carry out, and one of hot standby systems will use the reliable art of a large number versus a large number examined before "desire °: which can become active by the following methods (a) at meeting 12345, and it will broadcast a "desire" message. (b) Hear "desire message ' broadcast at meeting 12345. When the "desire ° message of that is heard [ else ], this succeeds as an active new data processing station. If succeeding it, an active new data processing station will tell that this takes over to all the user stations, and will announce which was the message properly processed by the last based on a global message sequence number. (C) In the case of the present invention, it is comparatively certain to throw away the message in the buffer from the side it is broadcast that a data entry station is. The recorder station specified by this time is because the response message is sent to a re-transmission station. Hub -- probably, one or the playback recorder station beyond it has received and stored the message certainly (even when a message cannot be received correctly, these can be searched from the recorder station which has this specified). The operation method is the same as a backup system (refer to Drawing 70). The same program can be used. 11 and online transaction processing -- high reliability -- To be -- in things and the past -- being reliable -- many data communications of opposite a large number were lacking -- a sake, the conventional system -- (Drawing 71 reference) -- typical -- a processing of an on-line transaction sake -- a large number -- an opposite -- the communication art of 1 is used. In communication of pair 1, processing is concentrated on 10 Casey John, and a majority of many users are connected to a single location. All the transactions must be processed by the single data processing station which accesses data. In this arrangement, there is a serious fault that a single data processing station will pay Okay. or all the processing loads. In the case of the present invention, the communication of a large number versus reliable a large number guaranteed for the on-line transaction can be used. For example, it should be considered rather than the application of pair 1 that many reservation systems [ a majority of ] of an airline are the things of opposite a large number. Many individual programs, such as an inquiry of an air fare, flight booking, fuel calculation, and a meal catering crew's schedule, are in the whole air line reservation system. Many users (for example, a travel agency and the manager of an airline) have to touch the most newest information. In the case of an existing air line data processing system, much art of pair 1 is used. All the fair flights, food, fuel, and crew information are inputted into a single data processing station (for example, large-scale mainframe computer). All the travel agencies and managers are connected to this single instruction single data stream station. A great processing load is applied to this single instruction single data stream station. - Especially, at the time of use of a peak, the data communication art of a large number versus reliable a large number guaranteed in this case can be used, and efficiency can be improved greatly. Many of un-indispensable programs are taken out from a single instruction single data stream station, and they can be carried out in the separated data processing station. Supposing one company of a travel agency does booking, the data communication art of a large number versus guaranteed reliable a large number is used for this booking information, and it can broadcast it to other stations of all the on a network. The individual program run from each of the data processing station which each station could maintain the database of itself and was these-separated can carry out [ newest ]-izing of each of these database individually. Many loading can be removed from a single instruction single data stream station. Thus, the present invention has big in bacto in the field of online transaction processing, The necessity for a large-scale mainframe computer of there being no delay unnecessary at the time of use of a peak, and having sufficient throughput to process a user's requirements is actually removed (instead). Newest-ization of a database is performed in the distributed form by all the stations on a network. The database which one of the questions asked by the expert of a database separated one time may be how how are maintained and especially maintained at a synchronization. The separated database is used by separated different program. Since it comes out, these are maintained independently. However, as the section of the backup system explained, a playback recorder can achieve the function as intensive backup. Since it is used by separated different station, these databases do not need to synchronize. However, it is guaranteed that all the inputs are received and processed by the same sequence. The design of this fact and the database distributed is simplified. The situation where a synchronization is needed is explained in the example (applications 12 and 13) of next application. The operation (refer to Drawing 72) by the application programmer in a data processing station is substitution by broadcast of transmission of the data to a central node. for example, : -- the substitution by CALL WRITE-CONFEREN CE of 5END data subroutine 12 The distributed database One of the greatest problems in the conventional distribution database system is locking of a remote database. Locking is used in order for another data-processing stage gin to prevent accessing one piece of data newest-ized. This lock process can cause a problem. And when a data processing unit causes failure of a lock, the complicated problem of processing of the locked data must be dealt with. In the present invention, the locking of a remote database can do 2 avoided completely. A re-transmission computer makes a newest-ized demand in-series efficiently, and all the newest-ized demands are broadcast. For example, in the air line reservation system distributed, there may be two duplication database data processing stations A and B. It is connected to data processing station A, and user X will assume that seat number 5 of flight 123 is required. This demand is broadcast. Suppose that it was connected to data processing station B, and user Y demanded seat number 5 of flight 123 in a similar manner by chance. This demand is broadcast similarly. However, these two demands are made in-series by the re-transmission station. When a demand [ A / for user X / data processing station ] is broadcast first, user X gets the seat. That the demand which calls data processing station B at a data processing station from the broadcast message was made in-series before the demand of it detects that the seat was got by user X. If doing it so, data processing station B must advise user Y with carrying out the book of another seat. Thus, although the complicated problem of locking of a remote database is avoided and a completely different alternative plan is provided in the present invention, this is more efficient than remote database locking art in fact. If the newest-ized message which can affect a participant called what 1000 is broadcast by chance, in order to avoid a problem, and in order to deal with the problem of a multiplex message transaction, a special "change check ° message can be used. A database newest-ized message can be received, without a general participating station changing that stored database with this art, until this receives the "change check" message broadcast from the newest-ized demand side station. Varnish Step 1 to which the following steps are concerned with operation of this concept by an application programmer (refer to Drawing 73) Database change request CALL WRIT E-CONFERENCE broadcast; Step 2 Broadcast message CALL READ-CONFERENCE reads; Step 3 Check of the message broadcast in order that the broadcast database change request might investigate whether other all are preceded with the piece of the data. When step 4 answer is YES, it is broadcast; step 5 about a "CONFIRMTo CHANGE' message by CALL WRITE-CONFE RENCE. When the answer to Step 3 is No (another database -) A station notifies a user of the failure which has required change on the piece of the same data. 13 Business information system The database had to compromise on the conventional system, in order to provide data input, newest[ real-time ]-izing, and an analysis function. In the past, since many data communications of opposite a large number were lacking, the conventional system was not able to run easily data input, newest[ real-time ]-izing, and an analysis program in the separated reliable data processing station. It is Being watched that a real-time analysis program changes frequently far rather than a data input program for the business needs of changing. Therefore, just after carrying out for the business needs which it is new or were changed, the necessity for design change may start a business information database. In the present invention, a database design can be simplified extremely. Newest[ real-time ]-izing and the analysis program can exist in a different data processing station. The inputted data is broadcast using the art of a large number versus guaranteed reliable a large number. The database for the formation of the real-time newest and analysis may completely differ from the thing about input data. Even if there is change to the analysis program based on the needs of business, the database about an input is not affected. The present invention gives big impact to the design of a database in this way. And it is double, there is no danger that the double information in a database which a double database can design by/or the guaranteed reliable message processing for newest-izing partially, and is different will become asynchronous. As for the following, the large-scale company which is an example of a business information system and : Existing has many products. Another manager is in each product. As for; product manager B to whom only product AE gets interested, product manager A gets interested in product B. These two managers have their own analysis program at a different analysis station. :(refer to Drawing 74) (a) data ON dregs Tation which can design a business information system as follows is programmed to broadcast data in the guaranteed reliable format of passing a secret conference which uses much art of opposite a large number, and is defined a priori. (b) Product managers A and B choose their own product from the broadcast message, and can analyze at the analysis station which self separated. They can change their analysis program, without affecting a company database. It can even perform having two different versions of their analysis program at two separated analysis stations. One of these is experimental. the operation by Maher is CALL WRITE-CONFERENCEI and l: so that :analysis station which it is as follows and obtains can take up data -- it is substitution about 'SENDdata subroutine. 14, a process control system The control system of the conventional process contains the master data processing station which receives an input from various kinds of data processing stations of other typically. this master station -- which -- others -- it judges whether a data processing station needs input, and sends input to these on 1 to 1 base. In this system, there is a fault that a powerful master data processing station is needed. in the case of the present invention, processing control was guaranteed -- being reliable -- much art of opposite a large number can be used -- 1 -- there is nothing about the necessity for a powerful master data processing station -- Si, It guarantees that distribute processing among many processing stations, then all the stations are newest-ized in addition in the present process control parameter and an input. for example, in manufacture of an alloy, I will be as follows -- convenience store -Tar [ for controlling heating of :(a) furnace ] A; (b) -- it can set in the various manufacture cycles computer C0 for controlling the sending rate of computer with temperature sensor B; for detecting the temperature of the alloy in the portion of a different furnace, or (C) several-kinds component -- the computer of these needs to communicate mutually. For example, when the temperature detected by computer B is too low, it is necessary to tell computer A about this, it needs to tell strength and computer C about heating, and needs to reduce sending of various components, or it is necessary to stop it. Only by computer B broadcasting the data, in the case of the present invention, computers A and C receive the data certainly, and can expect it to take proper action. Computers A and B. C may belong to another maker and does not have the necessity for the powerful master computer for the relay of information, either. operation (refer to Drawing 75) in computer B -- : -- substitution of S E N D data to CALL WRITE-C ONFERENCEl and master convenience store -Tar to : depend is included. In the parallel processing system conventional by substitution according [ the operation in convenience store -Tar A and C which hit receiving the broadcast information ] to :CALL READ-CONFERENCE, the portion which performs the processing task with two or same data processing station beyond it or from which it differs of the overall data-processing tasks can be performed - in order to carry out for obtaining, great organization and cooperation were required. In order to adjust the portion from which a parallel processing system differs, it is typical that specialization hardware and software are needed. Sadly, under the environment with which the maker mixed such specialization software and hardware, program, change, extension, or use could not be performed easily. In the case of the present invention, parallel processing can be performed using the art of a large number versus guaranteed reliable a large number. Supposing a certain input messsage is broadcast, one or the data processing station beyond it can process the input messsage simultaneously. For example, a large-scale database is manageable by a very powerful single instruction single data stream station. - Or the same database is manageable by the small-scale data processing station which could be divided into the portion of 1.000 or 1000 which is 1/1000 of the power of a single instruction single data stream station separated. It seems that the latter arrangement can process many transactions far [ selection Or ]. - Many of these small-scale data processing stations are because a parallel and different demand can be processed. the present invention -- parallel processing -- an application -- when used with cage gin, use under program change, extension, and maker mixture environment is made easy. An actual example (refer to Drawing 76) stands on the role which clarifies parallel processing application. right: <a application is an on-line patent search. In one or more user stations, search or reading can do the information on a patent number via a keyboard search. A database can be divided into much "volume ''. Each "volume ' includes the information based on a patent number which was able to be set in order. (b) When user station A desires information number 123456, the demand is broadcast using the art of a large number versus reliable a large number. I will assume that a volume 25 patent-database station is what has memorized the information about patent number 123456. It answers a demand. Other patent database stations of all the disregard the demand. while the volume 25 patent-database station has answered the demand [ station / user / A ], another patent database station can answer simultaneously (or -- being parallel) at another demand [ station / B / user phi ]. (C) When user station A wants to perform the keyboard search about a "photon", :(c 1) user station A whose step is as follows sends a keyboard search request message via the guaranteed reliable art of a large number versus a large number. all the patent database stations start a search simultaneously (or -- being parallel). That in which Finding spoiled "photon ' sends no response in the database. "photon -- " -- what found out the related reference uses the reliable art of 1 to 1, and answers user station A. (c2) Volume 3 patent database will leave other things, and will assume it as It was witnessed. One or more volume 3 patent-database stations can set on a network. These are called volume 3A, volume 3B, and volume 3C. (C3) A volume 3A patent database station broadcasts the message of having volunteered for this answering user station A, at meeting 12345, before answering user station A. A volume 3A patent database station hears the broadcast message on meeting 12345, in order to see this whether be the 1st volunteer from it. When an answer is No, a demand [ station / user / A ] is disregarded. When an answer is Ye s, using the guaranteed reliable art of 1 to 1, this answers a demand [ station / user / A ]. Thereby, the alternative duplicate of the portion of a certain database is attained. This can be called big remarkable progress in database technology. If it becomes like this, a database can be alternatively adjusted, in order to A fist to a user's request using the present invention. It is possible similarly to have one or a catalog station beyond it on the same network. These catalog stations supply Liszt based on Tobitsuku, a title, an artificer, a patent attorney, etc. who classified into beforehand. the remaining things of a patent database, simultaneously (or -- being parallel) execution are possible for these catalog stations. ; which the operation by the application programmer on a user station demands by varnish Step I CALL WRITE-CONFERENCE -- and Data is read by CALL READ-CONFERENCE after that [ step 2 ]. ; in which the operation step in a patent database station receives a demand by varnish Step I CAL L READ-CONFERENCE -- a result is returned to a demand user station via the art of Step 21 to 1. the example of application about the present invention of the above-mentioned statement -- although the great advantage is clearly shown to the message processing art of precedence, these examples are by no means restrictive. If it is reliable one-pair a large number and/, or the application that receives much profits from data communication of opposite a large number, which can receive profits from the present invention. Application called the company announcement of un-indispensable news transmission or a repetition may go into the category of the broadcast / multicasting transmission which is not a problem with guaranteed serious reception. however, about almost all the applications in connection with communication between one or more data processing stations, system operation is important for reliable message processing -- and -- actually -- 1 -- it is one serious parameter. Simulated broadcast The present invention is applicable also to a network without original broadcast capability. For example, in Drawing 82, a network comprises the node connected by the link of l versus l which a large number separated. The re-transmission station is arranged in position X which other nodes of all the can access from hop of the number of minimums at this. The recorder station specified can be arranged to all of an adjacent node like Y. Each node meant in order to send a broadcast message sends a message to the re-transmission station of X. via other nodes, if this message does not have course attachment Re, it may not become. If a message arrives at a re-transmission station, a global message sequence number and other time stamp information will be added. This raised message is transmitted to all the nodes which adjoin a re-transmission station from it. The node which adjoins these confirms whether this raised message is received before. In No, this raised message is sent at all those adjacent nodes other than the node sent in this. No transmission is carried out when it is what is received before. This process is repeated until all the nodes receive this raised message. This art is often known as message flooding. Reliability should be careful of being attained on all the links of 1 to separated 1 with communication art of 1 to 1 understood well which was indicated to 2 pages. About a secret conference, an informer sends a message to a management station first instead of sending to a re-transmission station directly. This art is called the broadcast by which the Shimuyu route was carried out. Although described in connection with what is considered to be an at present most practical and desirable example about the present invention, he should understand that it is not what is limited to the example the present invention was described to be. It has intention of covering the soul of the accompanying claim, various kinds of change included in within the limits, and an equivalent arrangement contrary to it. Doo and 1 Tower 78~bar 18FIG, I FIG, 2 Network B J-18 FIG, 3 FIG, 4 FIG, 5 FIG, 6 Special table 1004-502990 (49) FIG, 7 FIG, 8 FIG, 9 FIG, 10 Special table 1004-502990 (50) FIG, 11 FIG, 12 FIG, 13 FIG, 16 FIG, 17 FIG, 18 Patent Publication Heisei 4-502990 (52) FIG, 19 FIo, 20 FIG, 21 FIG, 24 FIG, 25 FIG, 28 FIG, 29 FIo, 32 FIo, 33 FIG, 34 FIG, 36 FIG, 39 FIG, 40 FIG, 43 FIG, 44 FIG, 47 FIG, 48 FIG, 51 FIo, 52 FIG, 55 FIG, 56 FIG, 57 FIG, 58 FIG, 59 FIG, 60 FIG, 61 FIG, 64 FIG, 65 FIG, 70 FIG, 71 FIG, 72 FIG, 74 FIG, 76 FIG, 77 FIG, 78 FIG, 79 FIG, 82 International search report mwmm-^Mission - and 1 A moon ON Ju Lu rPostcode Ha 1OnlyQlnBeam3^zeta