Memory buffer reserving system

[Detailed Description of the Invention]

The present invention relates to the data transmission system in a distributed shadow data exchange system, and relates to the data buffer request-to-print-out-files method at the time of data reception at Confrontation. The communications system O demand which is in Near city and the real time (real time) by which = Nambi-Tae Honpo IIK representation is carried out in reliability 011Ili, and performs data communications is increasing. The data exchange system of the packet exchange system is developed that this demand should be satisfied. The example of composition of a distributed data exchange system is shown in Drawing 1 as an example of this data exchange system. Comprising subsystem A, B, and C which stores the subscriber's line of 9 distributed data exchange system PCFi plurality, an interexchange channel, etc. in a figure, each subsystem A, B, and C performs data exchange processing independently with terminal (it is called member below) 8UB1~8UB4 and other data exchange offices. it is possible to extend a system easily by adding the Reason Peng subsystem with which this distributed data conversion system is adopted, and it comes out, and is -- a certain subsystem -- an obstacle -- it is alike in if an intermediary is not influenced by other subsystems, either but reliability high as a system is acquired. Drawing 2 shows the example of a loose communication sequence to a distributed data exchange system. The case where data communications are performed between member 8UB A village of Drawing 1 and member 80Bl is made into an example, and 2 dispatch side member 80Bl is subsystem A, the protocol (protocol) in which it has functions, such as a setup and release of A call, a stop, resending of data transfer, between subsystems according to X by CCITT advice publicly known between a member and a subsystem, and 25 procedures when sending data to mail arrival side member 8UB1 via B -- also following -- it is referred to as O. In subsystem AK, in the 28th, member 8UBm sends out Calling approval Cm (Cal I Request) first, Mail arrival directions CN (Incaning Ce1l.) are taken out with subsystem B to member 5UB1 by the side of mail arrival, Subsystem B receives zero display (Cal I Accepted), such as Response from member 8UB*, and Gills dispatch side member 8UB The KJI Success completion CC (Cal l Coanected) is sent. It is data DT- henceforth. The exchange of DTI, DT mind, and check number RR (Bccieve Ready) is performed. After member 5tnsi receives data DTl temporarily now, if the state where succession data is unreceivable occurs, receiving improper RNR (ReeieveNot Ready) will be returned. It is discarded in subsystem B at the time of that data reception that it is after dispatch side subSystem already receives the time of the next data reception from member 8UBs and sends out to subsystem B at this time, and Acknowledgement BR is not taken out. I Acknowledgement RRt subsystem B receives that data reception is possible, and at the time of a friend, member 8UBt gives the resending demand of data DT Red wine to subsystem people, and resumes henceforth usual data transfer. Sending [ in the end of data transfer / member SUBM ]-out-cutting demand CQ (C1earRequest) K Yo Si member 8UB1 receives cutting directions CI (C1ear Iadication), It completes by sending out Disconnection 1ICF (C1cJr Conf 1nnatian). Then, when the data length (the maximum user data length) which the member stored in a subsystem can receive differs by the dispatch side 2 arrival side, it is at the time when the maximum user data length of the member of 1 arrival side is longer than that of the dispatch side member. In order to unite with the maximum user data length of the mail arrival member of a self-subsystem the short paddle data block received from the partner subsystem in the subsystem which stored the member of 2 arrival side, to edit as one data and to transmit, A short paddle data block is accumulated in the memory buffer in a 1 Responsibility subsystem, and after receiving all the data blocks, it transmits to an applicable member as one data. The sequence which accumulates and edits this data block and transmits is shown in Drawing 3. In a figure, member 8UBB is subsystem B and subsystem At. - When passing and carrying out 2 arrival member 8UBA Heta transmission, it is To Si return it about the exchange of data 1T-~DTn and Acknowledgement RR, The memo Rebut 7 Achievement data of subsystem A is stored, and with 0, therefore subsystem A which send member 8UBgamma-Tha DT- (to -) by all the completion of data C reception, whenever it receives data DTo~DTn, a Sled buffer is caught, and it stores. However, when memory buffers run short while catching this memory buffer and assembling data, the data reception from a partner subsystem becomes impossible, a data receiving section layer will be interrupted, and the data block of the waiting for 2 assemblies will stagnate. If the shortage of memory buffers occurs in a plurality of Call processing, it will be in a Radiation state. The sequence which shortage of this memory buffer generated is shown in Drawing 4. this Drawing 4 is the same as the case where it becomes A state [ that it is unreceivable in the data reception of Drawing 2 ] 11, and it comes to be alike of Drawing as if the resending demand of data is given to the dispatch side subsystem at the time of Running which the memory buffer used by others becomes an opening, and twists as the mail arrival data reception of Sound is possible. Thus, 9 distributed data exchange system is extendibility. While excelling in reliability, there is a possibility of producing stay of data (message) in a subsystem in communication between the members from whom the maximum user data length differs, and it causes the fall of processing efficiency of subsystems, such as a data assembly, surveillance of a buffer, resending demand processing, etc. by data stay. That this problem should be solved, in 9 distributed data exchange system, there is the object of the present invention in providing the memory buffer request-to-print-out-files method which prevents stay of the data (message) based on the shortage of a memory buffer in the 9 arrival side subsystems, when it is communication between the members from whom the maximum user data length differs. To achieve the above objects, set the 9 present invention to the distributed data exchange system which comprises a plurality of subsystems. The data length received with the above-mentioned subsystem is a short paddle case from the permission receiving data length of a mail arrival terminal, When you transmit data to the above-mentioned arrival terminal and you cannot catch the memory buffer for the above-mentioned permission receiving data important point after carrying out request-to-print-out-files capture of the memory buffer for the above-mentioned permission receiving data important point at the time of top data reception and receiving all following data, suppose that reception is impossible. The 1 present invention is explained in detail based on an example below. Drawing 5 is a figure showing the communication sequence which takes the memory buffer request-to-print-out-files method of the present invention. In a figure, the example which transmits member 8UBA Heta from member 8UBB is shown, and setting release processing of a data link is the same as that of what is shown in Drawing 2, and it carries out a Ah Si abbreviation. Dispatch side member 8UBB sends to subsystem B as short paddle data (block) by the exchange of a series of data DT o =DT n and Acknowledgement RH. request-to-print-out-files capture ■ If data O which comes first from subsystem B to subsystem A first is sent, in subsystem A, a memory buffer will be carried out based on the maximum user data length of 1 arrival side member 8UBA. A memory buffer is released while assembling and sending out a series of data to one data DTo (o=n) at mail arrival side member 8UBA, when receiving data is received one by one after that and it receives to final data n1. At the time of (NR) which, on the other hand, cannot carry out request-to-print-out-files capture of the maximum user data important point part memory buffer of 2 arrival side member 8UBA when subsystem A receives data n+1 (new data), it reports that it is unreceivable to subsystem B, and is .. Succession data DTn+ The from mail arrival side member 5UBB' is refused as receiving improper RNR with subsystem B. Thus, invalid suspension of the receiving data in subsystem people, etc. can be eliminated, and processing efficiency can be raised. Drawing 6 is a processing lineblock diagram of the memory buffer request-to-print-out-files method of the present invention. For example, Drawing 6 explained in figures, it is data reception IF of subsystem A! It is inputted into analysis processing AN and the discernment information on data (for example, succession existence flag called M bit included in packet data) can distribute to either of data 3. final data 4 in the middle of initial-data 2. be [ maximum user data length ] alike corresponding to each terminal stored in a subsystem -- all the number TLB of memory buffers in the number NSB of the Si leg Was done required memory buffers, and a system -- as fixed data -- memory Ml, It is stored in MS and has same number USB [ as the number C8B of memory buffers ] of use memory buffers, and memory [ of which request-to-print-out-files memory buffer fi R8k storing is done ] M2.M3.M4 during the assembly corresponding to each terminal which serves as variable data by data-communications processing. The discrete value of 1 [ +] and CMP of G 1-04 are comparison machines a gate circuit, ADI, and AD2. The case where receiving data is [ following ] data and (C) final data in the middle of 0 initial data and (0) is explained. <-) When it is initial data, a subsystem receives the data to the mail arrival side member, and it is .. When it distinguishes that they are initial data (data sent first), it is an applicable member's required number N8B of memory buffers, Comparison machine CMP-% It inputs, To 5, the number 08B of use memory buffers, and the number of empty memory buffers that deducted the number of request-to-print-out-files memory buffers R8BF sum total from all the number TLB of memory buffers are inputted into CMP, and it is To 20, If not number[ of required memory buffers ]-securable, the processing which discards receiving data (message) and it makes impossible [ reception ] is started (9). If Jin in which the number reservation of required memory buffers is possible is got to know, gate Gl and gate sigma1 gate G4 will be opened (7). While buffer-quantity C3BK1 addition 6- Carrying out during an assembly, the number NBS of required memory buffers is applied to 888 request-to-print-out-files memory buffers, and it subtracts from (14) and the number R8B of request-to-print-out-files memory buffers one time, and it is 1 addition (11) f To to To 13 use buffer quantity U88. And reception continuation processing is started (8). In the There case [ Data in the middle of -> ] Homeland jlAN -- on the way -- receiving data -- 9 -- if things are understood -- (3), gate G1 -- good -- the salmon which opens G4 and adds C8B for the number of the memory buffers in 2 assemblies one time -- subtracting the number R8B of request-to-print-out-files memory buffers one time to 41 -- a To 13? use memory buffer (if it is identified that it is final data in -> final data reception (4)) It is the number CAB of memory buffers during a gate G4 * 0:00 difference assembly. It adds one and number of a1 request-to-print-out-files memory buffers R8 Bt-x subtraction is carried out, To t3L use memory A buff gamma number 08 B'il addition is carried out, gate Gs is opened to To 111 order, it assembles from the number NSB of required memory buffers, the number C8B of inside memory buffers is reduced, and it is the number of Me 7 Bach Noah of the Remaining. (-- the number of memory buffers which was not used -- "0 -- " -- there are also things --) -- it subtracts from request-to-print-out-files buffer quantity R8B, and clears the number CAB of memory buffers during 1 assembly. A memory buffer is released during the assembly which became unnecessary to the mail arrival side member about receiving data after sending out, and use buffer quantity is updated. Drawing 7 shows the explanation flow of the memory buffer request-to-print-out-files method of the present invention. The portion shown by A is the processing which is added by the present invention. If it is a discernment (22) L head about initial data when the short data reception (21) from the dispatch side subsystem occurs, an empty memory buffer will be caught (26). Place II in which an empty memory buffer gives a Age notice If (u) is performed and there is an empty memory buffer, request-to-print-out-files processing (27) of a memory buffer will be performed. It is a deed about the processing (24) which will assemble data and will transmit to a mail arrival member on the other hand if it is the completion of reception when it is succession data. If it is data on the way, waiting processing for 2 assemblies (25) will be performed. As explained above, according to the 2 present invention, it sets to a distributed data system, When the data received with the subsystem performs a data assembly by all the data reception rather than the maximum user data length of a mail arrival terminal in a short paddle case, Request-to-print-out-files capture of the number of memory buffers which balances the maximum user data length at the time of top data reception is carried out, Since the shortage of the K>In To memory buffers in data reception Tooru can be prevented and data maintenance management, such as a resending demand to invalid stay of data (message) and other subsystems, is made unnecessary, improvement in processing efficiency can be performed.

[Brief Description of the Drawings]

the 1st WJF! the [ the lineblock diagram of a distributed data exchange system, and ] -- the figure of the communication sequence taken two, and Drawing 6 -- lineblock diagram of the present invention. Drawing 7 is a processing 70-figure of the present invention. The number of memory buffers, 08B: The number of use memory buffers. B8B ; The number of request-to-print-out-files memory buffers, TLB; the total number of memory buffers Violet=20 The 3rd [2]