Communication method and system for the transmission of time-driven and event-driven Ethernet messages

Communication method and system for transmitting time-driven and event-driven Ethernet information

The present invention relates to a communication method used to transmit Ethernet information in a distributed real-time system. There are multiple network node computers, such as four network node computers. Each computer includes at least one communication controller, and by including one or The communication systems of the above communication channels are connected, and each communication channel is equipped with one or more smart star couplers.

Furthermore, the present invention relates to a star coupler used as a communication system for transmitting Ethernet information in a distributed real-time system, which includes multiple network node computers, such as four network node computers, each computer containing at least one communication Controller. The communication system includes one or more communication channels, network node computers are connected to each other through the communication channels, and each communication channel is configured with one or more smart star couplers.

The following documents are listed as follows: [1] US 5694542 published on December 12, 1989: Loosely coupled distributed computer system with real-time accurate node synchronization.

[2] EP 0 658 257 of December 18, 1996: Communication unit and method for transmitting information.

[3] US 5887143: Time-driven communication control unit and communication, published on March 23, 1999.

[4] AT 407 582 on June 15, 2000: An information distribution unit equipped with integrated protection devices to prevent crosstalk "fools" errors.

[5] AT 408 383 on March 15, 2001: A method and communication control unit for multi-original clock synchronization in a distributed real-time computer system.

[6] Austrian Patent Application No. 1723/2001 dated October 10, 2000: A method to accommodate slight deviations from specifications in a distributed, fault-tolerant, real-time system.

[7] Austrian Patent Application No. 429/2001 dated March 19, 2001: A communication method for establishing event channels in a time-driven communication system.

[8] The Institute of Electrical and Electronics Engineers (IEEE) Ethernet standard 802.3 at the Uniform Resource Locator (URL): HTTP://st andards.ieee.org.

[9] Kopetz, H. (1997). Real-Time Systems, Design Principles for Distributed Embedded Applications; ISBN: 0-7923-9894-7. Boston. Kluwer Academic Publishers.

[10] Sharon, O., Spratt, M., "A CSMA/CD compatible MAC for real-time transmission based on varying collision intervals" (CSMA/CD compatible MAC for real-time transmission based on varying collision intervals), in Yu: 1998 International Organization for Standardization Information Committee Meeting. The 7th Annual Meeting of the Computer and Communication Society of the Institute of Electrical and Electronics Engineers. Minutes of the meeting. Institute of Electrical and Electronics Engineers, No. 3, 1998, No. 3, pp. 1265 to 1272.

In the past two decades, the Institute of Electrical and Electronics Engineers (IEEE) Ethernet standard 802.3 [8] has been widely accepted, and because of the huge market for Ethernet controllers in the field of personal computers, the price of Ethernet-based communication systems has dropped significantly. Although the existing Ethernet protocol does not have good real-time characteristics, such as minimum uncertainty, due to price reasons, the use of Ethernet in real-time data processing is also increasing.

From [10], we can learn a CSMS/CD system in which information is divided into low and high priority. When the two pieces of information conflict, the high priority information will be processed first.

However, the real-time characteristics of the Ethernet protocol cannot be significantly improved by itself in the procedure proposed in this article.

An object of the present invention is to enable the transmission of Ethernet information to have good real-time characteristics.

This objective can be achieved using the methods of the categories mentioned at the beginning, in which the traditional Ethernet information (ET information) and time-driven Ethernet information (TT information) are distinguished according to the present invention. The TT information is known in advance between the transmitter and the receiver. When there is a time conflict between ET and TT information, the ET information involved in the conflict will be delayed or suspended, so that the TT information can be transmitted according to the constant delay time.

Furthermore, the purpose stated at the beginning can be achieved by using the star controller set according to the present invention described at the beginning to distinguish between traditional Ethernet information (ET information) and time-driven Ethernet information (TT information). The information is transmitted according to the constant delay time known in advance between the transmitter and the receiver. In this regard, when there is a time conflict between ET and TT information, the ET information involved in the conflict will be delayed or the transmission will be suspended. Constant delay time to transmit TT information.

Compared with the "no priority processing" solution disclosed in [10], according to the present invention, it is not necessary to wait for the transmission of low priority information to be completed, but low priority information will be suspended so that high priority information ("priority information") can be transmitted. Processing), so there is no need to wait for the longest running time of low priority information, and the constant waiting time can also be maintained within a short time.

By ensuring a constant delay time, it is possible to achieve a high degree of control engineering accuracy. The constant delay time is therefore of special importance. The reason is that according to the clock synchronization theory, the delay time changes (that is, the longest and shortest The difference in delay time) will reduce the accuracy of clock synchronization. In the clock synchronization algorithm, it is possible to consider the use of a known constant delay time, so that it will not affect the accuracy of clock synchronization. Since the granularity of the global time must be greater than the accuracy of clock synchronization, imprecise clock synchronization will result in a poor time reference. The granularity of the clock is rough, resulting in inaccurate time clarity of events. In addition, the change in the delay time also determines the accuracy of the synchronization of distributed actions in the distributed computer system.

The invention makes it possible to greatly improve the real-time characteristics of the communication system. This new communication system supports the parallel operation of event-driven and time-driven Ethernet messages in a single communication system. In the following, traditional Ethernet information is referred to as ET (Event Driven) information, and time-driven Ethernet information is referred to as TT (Time Driven) information. Time-driven information has constant delay time and minimum uncertainty.

The present invention brings the following major economic benefits. The lowest instability of TT information helps to develop a closed-loop control loop that highly controls the quality of the project. TT information helps to formulate high-precision global time. Global time supports the generation of accurate local time stamps when acquiring data, making it possible to improve the time specifications of the interface. In addition, traditional Ethernet controllers can be used without modification.

If the constant delay time is selected based on the output channel of the star coupler that can be cleared for transmission of the TT information to be received, the method of the present invention can be produced particularly simply.

In one embodiment, a designated segment of the information will indicate whether the information is a piece of TT information or an item of ET information.

Furthermore, the TT information may include a selective time segment indicating the moment of information transmission.

In this article, it will be a benefit if the time interval to be observed between two TT information transmissions is at least a constant delay time through pre-planning.

Furthermore, the above-mentioned purpose is achieved by using the system described at the beginning, that is, setting according to the present invention, distinguishing traditional Ethernet information (ET information) and time-driven Ethernet information (TT information), and according to the transmitter and receiver TT information is transmitted with a constant delay time known in advance. For this reason, when there is a time conflict between ET and TT information, the ET information involved in the conflict will be deferred or suspended, so as to ensure that the constant delay time Send TT information.

As mentioned above, if the selected constant delay time, the output channel of the star coupler can be cleared within the delay time for the transmission of incoming TT information, which will be a benefit.

In addition, it can be specified whether the indication information in a designated section of the information is a piece of TT information or an item of ET information.

Furthermore, the TT information may include a selective time segment indicating the transmission time of the information.

In this article, it will be a benefit if the time interval to be observed between two TT information transmissions is at least a constant delay time through pre-planning.

In a substantial embodiment of the communication system, it is stipulated that when the incoming information is TT information, it will be displayed to the star coupler through a piece of configuration information.

In this article, the star coupler distinguishes TT information and ET information, and transmits TT information according to a predetermined constant delay time. When there is a time conflict between ET information and TT information, it will stop balancing the ET information involved in the conflict. To ensure that the TT information can be transmitted with a constant delay time.

According to regulations, after the TT information is transmitted on time, the star coupler will retransmit the ET information that involves the balance and is suspended.

Furthermore, it can be specified that the star coupler uses the time segment contained in the TT information to calibrate its local clock.

In this article, if the star coupler uses the time segment contained in multiple TT information to calibrate its local clock in a fault-tolerant manner, it will be a particularly beneficial feature.

In addition, it can be specified that the star coupler is connected to a replicated star coupler in a group of network node computers, and the connection is carried out through a dedicated one-way channel on which all TT information transmitted by the star coupler All belong to the output form.

In addition, the star coupler can still check whether the TT information arrives within a pre-known time window around the transmission time based on its local time reference. When the star coupler receives a TT message early or late, it will The information is listed as an error message, so that all correct receivers will detect the information as an error.

The star coupler will decode each TT information and re-encode it based on its local time setting module.

The star coupler reads one or more selected sections of the TT information, and checks whether the contents of these sections meet the known standards that are pre-transmitted to the star coupler through the configuration information within the deferred time. If it does not match, the information is listed as an error message, so that all correct receivers will detect the information as an error.

In addition, according to the present invention, the communication controller uses the time section contained in the TT information to calibrate its local clock.

The communication controller calibrates its local clock in a fault-tolerant manner using the time segments contained in multiple TT messages.

In addition, when the transmission time indicated in the time zone of the information arrives, the TT information accepted by the application program executed by the network node computer will be automatically transmitted by the communication controller.

In addition, it is stipulated that the communication controller distinguishes ET and TT information, and the communication controller provides the ET information to the local application software that matches the semantics of the event. It is stored in the waiting queue as a new piece of information and is consumed by the application software. The communication controller provides the TT information to the local application software that conforms to the semantics of the state, and replaces the old information with new information, which is read by the local application software in a non-consumable manner.

Finally, the communication controller has two or more communication channels, which provide multiple identical copies of the same TT information. If a valid TT information is received on time by at least one of the redundant channels, the communication operation is regarded as success.

In the following, the present invention will be illustrated with reference to the diagrams. The diagrams shown are as follows: Figure 1 is the structure of a distributed computer system with a star coupler, and Figure 2 is the structure of a distributed computer system with two star couplers. 3 is a standard specification of the general Ethernet information structure, Figure 4 is a standard specification of the extended Ethernet information structure, Figure 5 is a TT Ethernet information structure, and Figure 6 is the TT Ethernet information structure A bit array of the TT parameter section.

In the next section, an embodiment of the new method is shown through an example, in which there are four network node computers connected by two repeating star couplers.

Figure 1 shows a distributed computer system together with a star coupler. The system includes four network node computers 111, 112, 113, and 114, each with a communication controller 121, 122, 123, and 124. Each of these controllers is connected to a two-way communication channel, and it includes a communication channel 109 The communication system is connected. A smart star coupler 101 is provided in this communication channel for central control communication. The star coupler 101 can be preset and observed through a selective independent communication channel 141.

Figure 2 shows a distributed fault-tolerant computer system with two star couplers. The system includes four network node computers 111, 112, 113, and 114, each with a communication controller 121, 122, 123, and 124. Each of these controllers is connected to two two-way communication channels, and each of these communication channels is connected to A smart star coupler 101 and 102 which perform central control of communication are connected. The star coupler 101 can transmit its information to the star coupler 102 through the channel 151, and can be preset and observed through the independent communication channel 141. The star coupler 102 can transmit its information to the star coupler 101 through the channel 152, and can be preset and observed through the independent communication channel 142.

Figure 3 shows the structure of general Ethernet information in accordance with the standard specification of [8]. After the 7-byte preamble 301, there is a start delimiter section 302, a destination address 303, a source address 304, a message length or message type 307, a variable data section 310, and an extension of the short message Optionally add a data section 311 and a frame check sequence 312.

Figure 4 shows the structure of the extended Ethernet information according to the standard format of [8]. In addition to the sections described in FIG. 3, the section 305 is provided with an identifier of the extended information, and the section 306 is provided with a label type section. In the tag type section, the user can determine the priority of the information, and the highest priority can be used to identify the TT information according to the present invention. The above-mentioned identifier conforms to the Ethernet standard [8]. It must be pointed out that in the Ethernet standard, the coding capacity of section 305 has not been fully used, so this section can also be used to identify TT information.

Figure 5 shows the structure of TT Ethernet information. In addition to the sections described in FIG. 4, the TT parameter section is introduced in section 308, and the selected transmission time of TT information is indicated in section 309. Standard Ethernet controllers available on the market read user-specific data sections in sections 308 and 309. In the TT parameter section 308 is information about the structure and type of TT information.

FIG. 6 shows the content of the bit array of the TT parameter section 308. If this bit is set in section 601 (low order bit), it means that the transmission time in section 309 is contained in the TT information. If this bit is set in field 602, it means that the information comes from a transmitter with accurate clock time, which can be used for clock synchronization.

For example, a network node computer (for example, 111) intends to transmit a piece of TT information, and it sets the code of the TT information in the information section 306 and transmits the information. Another option is that the application software running on the network node computer can set the information in the information setting bit 601 and write the intended transmission point in the information section 309, and then it can be accurately set by the Ethernet communication controller. The transmission time 309 is set to start transmission autonomously. If the transmitter sets the information bit, the information will carry a particularly accurate time indication, which can be used for clock synchronization of other controllers.

The star coupler analyzes the incoming information and uses the section 306 to determine whether it is a TT message or an ET message that is about to arrive. In the case of TT information, the star coupler determines the intended output channel based on the section 303, for example, the node 114 in FIG. 1. If an ET message is directly transmitted on this channel, the star coupler will immediately suspend the transmission operation and clear the channel to the node 114 within a known constant delay time for further transmission of the newly arrived TT message. The selected delay time must be sufficiently long, so that in each case the output channel can be cleared within this delay time for the transmission of TT information. In the pre-planning of TT communication, it is necessary to ensure that the interval between TT messages is greater than the delay time. In individual cases, the star coupler accurately observes the constant delay time between the beginning of receiving a TT message and the beginning of the transmission of a TT message. If the star coupler has stopped transmitting the conflict-related ET information, it can retransmit the suspended ET information after transmitting the TT information on time. The star coupler can also take on the maintenance function as described in [4] to detect and separate error messages, thereby preventing error propagation. If an item of TT information in the section 309 obtains the transmission time, the star coupler can check whether the information arrives within the known tolerance interval about the transmission time according to [6], otherwise the information will be rejected. Another option is to use pre-transmitted configuration information through channel 141 to convey the input channel and input time that are expected to receive TT information to the star coupler. This information redundancy in a fault-tolerant system can prevent a faulty computer node from entering incorrect transmission time. Since the star coupler encodes the output information based on its own oscillator and its own power supply, the SOS error transmission from the transmitter to the receiver [4] will be stopped. The star coupler can first calibrate its local clock according to the measurement of the TT message start receiving time, and set its clock so that it will receive the global time value 309 at the time of reception, which is contained in the message [5].

A continuous fault-tolerant clock synchronization can be realized as follows: According to each synchronization information marked by the segment 602, the star coupler will determine the reception time of the synchronization information measured by its local clock and the transmission carried in the information segment 309 The interval between time [5].

This interval measures the deviation between the receiver clock and the transmitter clock. If multiple pieces of such information exist, it is possible to use known fault-tolerant synchronization algorithms to calculate the correction factor of the star coupler clock, as described on page 61 of [9]. This fault-tolerant synchronization method can also be implemented in the hardware of the star coupler [1]. In a fault-tolerant system [2, 3], if there is a repeated communication channel consistent with Figure 2, each star coupler can pass a dedicated connection channel (the star coupler 101 is channel 151, the star The coupler 102 is channel 152) to transmit all TT information to another star coupler, so that when the latter itself has no input information, its clock can also be calibrated. In a fault-tolerant system, the star coupler can check the information data section 310 according to the criteria for reporting the configuration information to it within the deferred time to detect the data section of the transmitter. The information detected as an error will not be transmitted by the satellite.shapecoupler transmission.

If the receiving communication controller finds the transmission time in the section 309 of the arriving TT message, it can calibrate its local clock by measuring the time when the message is received, and set its clock so that it will receive the message at the time of reception. The global time value 309 contained in plus the constant delay interval caused by the star coupler [5]. Continuous fault-tolerant clock synchronization can be achieved as follows: According to each synchronization information marked by section 602, the communication controller will determine the interval between the reception time of synchronization information measured by its local clock and the transmission time contained in section 309. This interval It will be shortened according to the known delay interval of the star coupler. This shortened interval measures the deviation of the receiver clock from the transmitter clock. If multiple pieces of such information exist, it is possible to calculate the correction factor of the star coupler clock using a known fault-tolerant synchronization algorithm as described on page 61 of [9]. This fault-tolerant synchronization method can also be implemented in the hardware of the receiving communication controller [1]. For example, the application software of a computer node (for example, 111) enters the intended information transmission time in the information section 309, and the extended communication controller according to the present invention can automatically initiate transmission at the correct transmission time [2, 3] ]. In the interface between the receiving communication controller (such as 121) and the application software, the expanded communication controller can provide ET information and TT information respectively. ET information generally contains information about the event, and must be processed in accordance with the semantics of the event [7]. The event semantics stipulates that the arriving information must be temporarily stored in the waiting queue and sent to the user processing program exactly once. TT information generally includes status data, which can be provided in accordance with the status semantics in the public storage. Receipt of new TT message will overwrite the stored value of old TT message with the same name. The receiving program will read the status data in a non-consumable way. In a fault-tolerant system that provides multiple multiple independent communication channels, for example, through two channels as shown in Figure 2, information will be transmitted in a repeated manner. In this type of system, if at least one of the duplicate information copies arrives at the receiver, the communication is successful.

Finally, it must be noted that the foregoing specific embodiment of integrating time-driven and event-driven information in the Ethernet is only one of many possible implementation variations of the present invention.

For example, the decision about whether a piece of information arriving at the star coupler is a piece of TT information may not be made based on the information content in section 306 or section 305, but based on the time when the star coupler received the information. Made. In this case, it is necessary to report to the star coupler in advance through the item configuration information when and on which channel the expected TT information will arrive.

This can also be used for communication controllers.

A basic feature of the present invention is that the currently commercially available Ethernet controller can transmit and receive time-driven information without modification.