Communication method and system for transmitting timed and event-driven ethernet messages

The invention relates to a communication method for transmitting Ethernet messages In a distributed real-time system in which a plurality of node computers, Eg four node computers, which have at least one communication controller, Via a communication system consisting of one or more communication channels , Wherein in each communication channel one or more intelligent Star coupler, and wherein between conventional Ethernet messages (ET messages) and time-controlled Ethernet messages (TT messages) , And the TT messages with a constant delay time known a priori Between transmitter and receiver.

Such a method is known from WO 01/095550 A.

Furthermore, the invention relates to a star coupler for a communication system for Transmission of Ethernet messages in a distributed real-time system comprising one A plurality of node computers, for example, four node computers, which are connected via at least one communication controller Wherein the communication system consists of one or more Communication channels, via which the node computers are connected to one another , And wherein, in each communication channel, one or more intelligent star couplers Are arranged.

In the following text, reference is made to the following literature:<sl><li>[1] US5694542</li><li>[2] EP 0 658 257</li><li>[3] US5887143</li><li>[4] AT 407 582</li><li>[5] AT 408 383</li><li>[6] Austrian patent application 1723/2001 "Method for tolerating slightly-offspecification Error in a distributed fault-tolerant real-time system. " 10.10.2000 </li><li>[7] Austrian patent application 429/2001 "Communication procedures for realization Of event channels in a time-controlled communication system. " on March 19, 2001</li><li>[8] IEEE Ethernet Standard 802.3 at URL: http://standards.ieee.org</li><li>[9] Kopetz, H. (1997). Real-Time Systems, Design Principles for Distributed Embedded Applications; ISBN: 0-7923-9894-7. Boston. Kluwer Academic Publishers</li><li>[10] Sharon, O., Spratt, M., "A CSMA / CD compatible MAC for real-time transmission based On varying collision intervals ". In: INFCOM '98, Seventh Annual Meeting Joint Conference Of the IEEE Computer and Communications Societies, Proceedings. IEEE, Volume: 3, 1998, Pages 1265-1272 vol. 3.</li></sl>

In the past twenty years, the IEEE Ethernet standard 802.3 [8] has gone so far That due to the existing mass market for Ethernet controllers in the Personal computer sector the cost of Ethernet-based communication systems greatly Have fallen sharply. For these reasons, Ethernet is also becoming increasingly popular in the industry Real-time data processing, although the existing Ethernet protocol does not Good real-time characteristics, such as minimal jitter.

From [10] a CSMS / CD system is known, in which messages in such lower and High priority, wherein in the event of a conflict between two messages of said message With high priority.

However, with the method proposed here alone, the real-time characteristics can be achieved Of the Ethernet protocol.

It is an object of the invention to provide the transmission of Ethernet messages with good Real-time characteristics.

This object is achieved by a method mentioned at the outset in that, according to the invention, In case of a temporal conflict between ET and TT messages the transport The conflicting ET message is canceled to receive the TT message with the Constant deceleration time.

Furthermore, the above-mentioned object is achieved with a star coupler mentioned at the outset Which, according to the invention, is configured to switch between conventional Ethernet messages (ET) messages and time-controlled Ethernet messages (TT messages) The TT messages with a constant delay time known a priori Between the transmitter and the receiver, wherein in the case of a temporal Conflict between ET and TT messages the transport of the conflicting ET message Is aborted to the TT message with the constant delay time Transport.

In contrast to the "non pre-emptive" solution disclosed in [10], The end of the transmission of a low priority message has not been awaited , But the lower priority message is canceled High-priority message ( "pre-emptive"). This is not the case either It is necessary to wait for the maximum duration of low priority messages, and The constant latency can thus also be kept short.

Also in WO 01/095550 A such a solution according to invention is not shown, there The transport of the data stored in the Conflicting ET message is only delayed.

By ensuring a constant deceleration time, a high control technology can be achieved Accuracy can be achieved. The constant deceleration time is therefore for the reason Of particular importance since, as is known from the theory of clock synchronization, The variability of the delay time (which is the difference between the maximum and the Minimum delay time) degrades the accuracy of clock synchronization. An a priori known constant delay time can be found in the clock synchronization algorithm And therefore does not affect the accuracy of the Clock synchronization. An inaccurate clock synchronization leads to a bad clock synchronization Time basis, since the granularity of the global time must be greater than the accuracy of the Clock synchronization. A rough granularity of the watches results in an inaccurate temporal Resolution of events. Further, the variability of the delay time is determined Also the accuracy of the synchronization of distributed actions in a distributed computer system.

The present invention enables the real-time characteristics of an Ethernet-based device Communications system. This new Communication system supports the parallel operation of event - controlled and time - controlled Ethernet messages in a single communication system. Hereinafter The classic Ethernet messages are used as ET (event-triggered) messages and the Time-triggered Ethernet messages as TT (time-triggered) messages. The TT messages have a constant delay time and minimal jitter.

The invention provides the following significant economic advantages. The minimum Jitter of the TT messages makes it possible to set up control loops of high control technology Quality. TT messages allow you to set up a global time With good precision. The global time supports the generation of accurate local timestamps Pel in the data acquisition and allows the temporal specification of the interfaces to improve. In addition, conventional Ethernet controllers can be without any changes Can be used.

The method according to the invention can be implemented particularly simply if the constant Deceleration time is selected such that within this delay time the output channels Of the star coupler is cleared for the transport of the incoming TT message Can be.

In one embodiment, in an excellent field of the message, Whether the message is a TT message or an ET message.

Furthermore, an optional time field can be contained in a TT message, which is the transmission time Of the message.

It is of advantage, if an a priori planning is already specified that Between the transport of two TT messages a time interval of at least Of the constant deceleration time.

Furthermore, the above-mentioned object is achieved with a system mentioned at the outset, According to the invention, is configured to switch between conventional Ethernet messages (ET) messages and time-controlled Ethernet messages (TT messages) And the TT messages with a constant delay time known a priori Between the transmitter and the receiver, wherein in the case of a temporal Conflict between ET and TT messages the transport of the conflicting ET message Is aborted to the TT message with the constant delay time Transport.

As already mentioned above, it is advantageous if the constant delay time is selected in this way Is that within this delay time the output channels of the master coupler for The transport of the incoming TT message can be cleared.

It can also be provided that is marked in an excellent field of the message Is whether the message is a TT message or an ET message.

Further, in an TT message, an optional time field may be included, which is the transmission time Of the message.

It is of advantage in this case if an a priori planning already stipulates that between The transport of two TT messages a time interval of at least that Constant deceleration time.

In a specific embodiment of the communication system, it is provided that To the star coupler via a configuration message to which Time points are incoming messages of TT messages.

The star coupler distinguishes between TT messages and ET messages and Transports the TT messages with a constant delay time known a priori, And if there is a temporal conflict between ET and TT messages, he breaks the transport Of the conflicting ET message off to the TT message with the constant delay time Transport.

In this case, it is then provided that the star coupler, after the timely transmission of a TT message retries the conflicting terminated ET message.

It may also be provided that the star coupler transmits its local clock, using the Of a TT message.

It is particularly favorable when the star coupler transmits its local clock by using the Several TT messages.

In addition, it can be provided that the star coupler has a dedicated unidirectional Channel on which all TT messages which the star coupler transports are output With the replicated star couplers within a cluster of node calculators connected is.

In addition, it is then still possible that the star coupler for each TT message on the Based on its local time, verifies whether the TT message is known within a priori Time window by the time of transmission contained in the message, and the star coupler If a TT message is received prematurely or late Mutilates that all correct recipients recognize the message as faulty.

The star coupler decodes each TT message and encodes it on the basis of its local Time.

The star coupler reads one or more selected fields of TT messages and checks During the delay time, whether the content of these fields with known criteria, the To the star coupler via a configuration message a priori. If a match is not given, the message is mutilated, That all correct recipients recognize the message as faulty.

Furthermore, according to the invention, it is also provided that the communication controller can control its Local clock is synchronized using the time field contained in a TT message.

The communication controller synchronizes its local clock using the multiple clocks TT messages are fault-tolerant.

In addition, the communication controller sends one from one on a node computer Executed application, the TT message will be autonomous as soon as the one in the message In the time field is reached.

Furthermore, it is provided that the communication controller switches between ET and TT messages And the communication controller responds appropriately to the ET messages The event emulation to the local application software, whereby a new Message is placed in a queue, from the application software And the communication controller reads TT messages accordingly The state semantics to the local application software, whereby a new Message replaces the old version of the message and reading it through the local application software Not consumed.

Finally, the communication controller has two or more independent ones Communication channels on which identical copies of a TT message are offered, And the one communication process considered successful, if at least One of these redundant channels arrives in time for a valid TT message.

The invention is explained in more detail below with reference to the drawing. In this shows<dl tsize="6"><dt>FIG</dt><dd>The structure of a distributed computer system having a star coupler, </dd><dt>FIG</dt><dd>The structure of a distributed computer system having two star couplers,</dd><dt>FIG</dt><dd>The establishment of a standardized normal Ethernet message,</dd><dt>FIG</dt><dd>The establishment of a standardized extended Ethernet message,</dd><dt>FIG</dt><dd>The structure of a TT-Ethernet message, and</dd><dt>FIG</dt><dd>The bit fields of a TT parameter field of the TT Ethernet message.</dd></dl>

In the following, an implementation of the new method is presented with a possible example with Four node computers connected via one or two replicated star couplers, shown.

FIG. 1 shows a distributed computer system with a star coupler. It consists of the four Node computers 111, 112, 113, and 114, which are connected via a respective communication controller 121, 122, 123, and 124 each having a bidirectional communication channel port, and the Via a communication system consisting of a communication channel 109 are. An intelligent star coupler 101 is located in this communication channel For the central control of the communication. The star coupler 101 can be connected via an optional Separate communication channel 141 can be initialized and monitored.

FIG. 2 shows a distributed fault-tolerant computer system with two star couplers. It Consists of the four node computers 111, 112, 113, and 114, which are connected via a respective communication controller 121, 122, 123, and 124 each having two bidirectional communication channel connections feature. Each of these communication channel connections is connected to an intelligent Star couplers 101 and 102, which are the central control of communication The The star coupler 101 can transmit its messages via channel 151 to the star coupler 102 and can be initialized via the separate communication channel 141 and to be observed. The star coupler 102 can transmit its messages via channel 152 to the Star coupler 101 and can be initialized via the separate communication channel 142 And be observed.

FIG. 3 shows the structure of a standardized normal Ethernet message according to [8]. After a Preamble 301 with a length of 7 bytes is the start Delimiter Field 302, the destination address 303, the sender address 304, the message length, and the type of the Message 307, the variable data field 310, the optional PAD field 311, to the short Messages are extended, and the frame check sequence 312.

FIG. 4 shows the structure of a [8] standardized extended Ethernet message. In addition to the fields described in FIG. 3, an identifier is found in field 305 For the extended message and field 306 a tag type field. In this tag-type Field, the user can set the priority of a message. The highest priority can be Can be used to identify a TT message according to this invention. A Such identification is conform to the Ethernet standard [8]. It should be noted that Ethernet standard the code allocation for the field 305 is not yet fully exploited and Therefore this field could also be used for the identification of a TT message.

FIG. 5 shows the structure of a TT Ethernet message. In addition to those described in FIG A TT parameter field is inserted in field 308, and field 309 is the optional field Sending time of the TT message specified. A standardized product available on the market Ethernet controller provides user-specific data fields in fields 308 and 309. In the TT parameter field 308, there is information on the structure and type of the TT message.

FIG. 6 shows the contents of the bit fields of the TT parameter field 308. If the bit in field 601 (low order Bit), this means that in the TT message, the transmission time in field 309 Is included. If the bit is set in field 602, this means that the message is from a Transmitter with a precise clock time and is used for clock synchronization Be allowed.

If a node computer, for example 111, wants to send a TT message, it sets in the message field 306 the identifier for a TT message and sends the message. Alternatively, the On a node computer, bit 601 in the message And write the desired transmission time into the field 309 of the message. Of the Sendstart can then be autonomous by means of an extended Ethernet communication controller according to the invention Precisely at the fixed transmission time 309. If the Transmitter sets the bit 602 of the message, then the message includes a particularly precise one Which are used for the clock synchronization of the other controllers Can

The star coupler parses an incoming message and provides field 306 Determines whether a TT message or an ET message is received. In the case of a TT message The star coupler determines the desired output channel on the basis of the field 303, eg To node 114 in FIG. 1. If an ET message is being sent on this channel, The star coupler immediately aborts this transmission process and makes the channel to node 114 Within a known constant delay time Δ for the further transport of the Just incoming TT message. The deceleration time Δ must be selected as long as this occurs In any case, the output channel within this delay time Δ for the Transport of the TT message. Within an a priori planning Of the TT communication, it must be ensured that the distance between consecutive TT messages is greater than the delay time Δ. In the individual case the Star coupler, this constant delay time Δ between the beginning of the reception A TT message, and the beginning of the transmission of a TT message. If the Starcoupler the transportation t a conflicting ET message has aborted, It can send the terminated ET message after the timely transport of the TT message send again. The star coupler can also have Guardian functions, as described in [4] Are able to detect and isolate faulty messages and thus isolate the Propagation of errors. When a TT message in field 309 indicates the transmission time , The star coupler can check whether the message according to [6] Within a known tolerance interval around the transmission time point and the Discard the message if it is not. Alternatively, the star coupler can be used An a priori transmitted configuration message can be communicated via channel 141 Which input channels are expected at which points in time TT messages. These Information redundancy in a fault tolerant system prevents a more faulty system A wrong transmission time in the message. Since the star coupler has the Based on its own oscillator and its own power supply Output, the transmission of an SOS error from the sender to the receiver [4]. The star coupler can initially synchronize its local clock by pressing Measures the beginning of the arrival of a TT message, and adjusts its clock to allow it to This time of arrival the value of the global time 309 contained in the message Would have accepted [5].

A continuous error-tolerant clock synchronization can be implemented as follows: The star coupler determines the synchronization message for each synchronization message marked in field 602 Interval between the arrival time of the synchronization message measured with its local clock And the transmission time [5] contained in field 309 of the message. This interval is a measure of the deviation of the clock of the receiver from the clock of the clock Station. If a number of such measurements are present, a known method can be used Error-tolerant synchronization algorithm, as described in [9] on page 61, the correction factor For the clock of the star coupler. Such a fault tolerant Synchronization method can also be implemented in the hardware of the master coupler [1]. In a fault tolerant system [2,3], in which replicated communication channels Corresponding to FIG. 2, each star coupler can be connected via a dedicated connection channel (151 for star coupler 101 and 152 for star coupler 102) all TT messages The other star coupler so that it can also synchronize its clock, if No message arrives at the own input. In a fault tolerant system, the Star coupler within the delay time Δ the content of the data field 310 of the message According to it, notified via a configuration message, To detect data errors of the transmitter. A message recognized as faulty Not forwarded by the star coupler.

When a receiving communication controller is in field 309 of the incoming TT message The transmission time, he can initially synchronize his local clock, By measuring the beginning of the arrival of this message, and adjusting its clock such that At this time of arrival, the value of the global time contained in the message 309 plus the constant delay interval .DELTA. Caused by the star coupler Would have [5]. A continuous error-tolerant clock synchronization can be as follows The communication controller determines at each of the fields marked in field 602 The interval between those measured with its local clock Arrival time of the synchronization message and that in field 309 of the message Time. It shortens this interval by the known delay interval Δ of the star coupler. This shortened interval is a measure of the deviation of the Clock of the receiver from the clock of the transmitter. When a number of such measurements , A known error-tolerant synchronization algorithm as described in FIG [9] on page 61, the correction factor for the clock of the receiver can be calculated. Such a fault tolerant synchronization method can also be used in the hardware of the Received communication controller [1]. If the application software Of a computer node, eg 111, into the message field 309 the intended Transmission time of the message, an extended communication controller according to the invention can be used The start of the transmission autonomously at the precise correct transmission time [2,3]. At the interface between the receiving communication controller, Eg, 121 and the application software, a communication controller which is expanded according to the invention ET messages and TT messages. ET messages Usually contain information about events and must be made after the Event semantics [7]. The event emotions require that incoming Messages are queued in a queue and sent exactly once to the User process. TT messages usually contain status data, Which are offered in a common memory according to the state semantics Can be. The arrival of a new TT message overwrites the memory value An older TT message of the same name. The receiving process reads status data Not consuming. In a fault tolerant system that has several independent ones For example via two channels as in FIG. 2, are messages Replicated. In such a system, communication is successful when At least one of the replicated message copies arrives at the receiver.

Finally, it should be noted that the above-described concrete implementation of the Integration of time - controlled and event - driven messages in Ethernet only one Of many possible implementation variants of the present invention. For example, it is possible to decide whether to arrive at the star coupler Message is a TT message, not the message content in field 306 or field 305 But from the time of the arrival of a message at the star coupler. In Such a case must be transmitted to the star coupler a priori via a configuration message When and on which channel a TT message is to be expected. The same applies to the communication controllers.

It is an essential feature of this invention that existing commercially available Ethernet controller can send and receive timed messages without modification can.