HK1143473A - Communication method and device for efficient and secure transmission of tt ethernet messages - Google Patents
Communication method and device for efficient and secure transmission of tt ethernet messages Download PDFInfo
- Publication number
- HK1143473A HK1143473A HK10109938.4A HK10109938A HK1143473A HK 1143473 A HK1143473 A HK 1143473A HK 10109938 A HK10109938 A HK 10109938A HK 1143473 A HK1143473 A HK 1143473A
- Authority
- HK
- Hong Kong
- Prior art keywords
- time
- information
- star coupler
- ethernet
- communication method
- Prior art date
Links
Description
Technical Field
The invention relates to a communication method for transmitting time-triggered Ethernet messages in a distributed real-time system, comprising a plurality of node computers, wherein each node computer is provided with at least one Ethernet controller, which is directly connected via data lines to ports of time-triggered star couplers explicitly assigned to the node computer, and wherein a plurality of time-triggered star couplers can be directly or indirectly connected to one another via one or more data lines to form a closed time-triggered network.
Furthermore, the invention relates to the use of a time-triggered star coupler for propagating ethernet information in the above-mentioned communication method.
Background
Over the past 20 years, the Institute of Electrical and Electronics Engineers (IEEE) ethernet standard 802.3[5] has gained wide acceptance, and the price of ethernet-based communication systems has dropped significantly due to the large market for current ethernet controllers in the personal computer field. For reasons of price, the use of ethernet for real-time data processing is also increasing. European patent No. EP 1512254 [4] discloses a method that makes it possible to transmit time-triggered information with good real-time characteristics within an extended ethernet system (hereinafter referred to as TT (time triggered) ethernet).
In TT ethernet [4], a distinction is made between two types of information, legacy ethernet information (hereinafter referred to as ET (event triggered) information) and new type TT information. The TT message is characterized by containing a bit pattern (bit pattern 88d7) of IEEE Ethernet Standard management [5] grant in the Ethernet class field. When ET messages (i.e., conventional ethernet messages) come from an open environment that is not coordinated in time and thus may have time conflicts with each other, all TT messages in TT ethernet are assumed to be transmitted on a predetermined schedule in the closed TT network without interfering with each other. The closed TT network includes a plurality of node computers communicating through one or more TT star couplers.
Since the time difference between two TT messages must be greater than twice the accuracy ii to be able to exclude any possibility of TT message collisions [2], the efficiency of the transfer of useful data of TT messages in the TT ethernet network [4] depends to a large extent on the accuracy ii [6] of the time synchronization of the TT ethernet controllers in the node computers. Hardware support [1] requires very precise time synchronization in the range of about 1 μ sec), which is not found in commercial ethernet controller products on the market. If time synchronization is performed in software, it is difficult to achieve an accuracy of more than 50 μ sec, i.e. to eliminate the possibility of TT information collisions in the TT network, a time gap of at least 100 μ sec must be arranged between two TT ethernet messages. If it is assumed that most of the TT information is transmitted in the 100 megabit (Mbit)/second ethernet system for a time much shorter than 100 μ sec, the efficiency of useful data may be much less than 50% when using commercial ethernet controller goods on the market.
Disclosure of Invention
The object of the present invention is to improve the rate of useful data and data security in a distributed real-time computer system in which node computers communicate information via time-triggered ethernet as disclosed in, for example, EP 1512254.
The present invention achieves this object by a network transmission time, such as a periodic network transmission time (which is pre-assigned to each TT ethernet message by a scheduler), wherein a port of the TT star coupler delays the arrival of a TT ethernet message from a node computer to the next network transmission time of such TT ethernet messages in its system time, and the port transmits the TT ethernet message into the TT network at that network transmission time or within a certain time period starting from the network transmission time upwards.
A distinction is thus made between the node computer transmission time (KNSZPKT) and the network transmission time (NWSZPKT) of the information. According to the invention, it is proposed to adapt the TT star coupler such that TT Ethernet information arriving from the node computer is delayed at the intelligent port of the TT star coupler to the NWSZPKT, in order then to be able to transmit it into the TT network exactly at the NWSZPKT. The KNSZPKT interpreted from the system time of the sending node computer must be timely pre-existing from the NWSZPKT so that, in any event (i.e., even if the system times of the node computer and TT star coupler are at the limit of the accuracy gap [6 ]), the start of the message reaches the TT star coupler at the time of the NWSZPKT interpreted from the system time in the TT star coupler. Between KNSZPKT and NWSZPKT, time and semantic checks can also be made on the arriving TT information through the intelligent port of the TT star coupler to improve fault detection. Since the node computer and the TT star coupler form two separate fault-accommodating regions [7, 3], the possibility of fault propagation of erroneous information can be reduced by examining the information in the independent TT star coupler. In addition, the intelligent port of the TT star coupler may encode information arriving from the node computer such that the TT information is transmitted in the network in encoded form.
Further advantageous embodiments are set forth in the dependent claims.
The present invention yields the following significant economic benefits:
even with the commercial use of ethernet controllers on the market, and time synchronization of the ethernet controllers using software, the useful data efficiency of TT information transfer can be increased by much more than 90%.
Examining the TT information within the intelligent port of the TT star coupler reduces the possibility of error propagation and makes diagnosis easier.
The information is encoded at the intelligent port of the TT star coupler, so that the safety of a real-time system is improved without adding extra burden to an application computer.
A conventional ethernet controller may be used without changing the hardware used to transmit the TT and ET messages.
Drawings
The invention is explained in detail below on the basis of the drawings which are non-limiting examples. In the drawings, there is shown in the drawings,
figure 1 shows the structure of a distributed computer system with a TT star coupler,
FIG. 2 shows the transmission time of information, an
Fig. 3 shows the structure of TT ethernet information.
Detailed Description
The following paragraphs show embodiments of the novel method in a possible example with three node computers connected by a TT star coupler.
FIG. 1 shows a distributed computer system with a TT star coupler 101, the system comprising three node computers 111, 112, 113 connected to the TT star coupler 101 by bidirectional lines, the TT star coupler 101 being connectable to other TT star couplers by lines 100, thereby being part of a Time Triggered (TT) network comprising a plurality of TT star couplers, the system time assumption for all TT star couplers making up the TT network is a time base [6] (better than 1 μ sec) with a common high accuracy, the TT star coupler 101 comprises three smart ports 121, 122, 123 which can exchange information through a real Ethernet switch 102, each smart port 121, 122, 123 has autonomous processing capabilities such that information can be received in parallel and processed at the port at the same time At a given port 121, TT Ethernet messages arriving from a node computer (e.g., node computer 111) are delayed at port 121 until the next periodically returned network transmission time (NWSZPKT) for such messages is reached in TT Ethernet [4], the type of message is explicitly set by the periodic transmission time of the node computer in the TT network.
Fig. 2 shows the chronological order of sending the periodic time trigger information in a periodic loop demonstration, taking as an example the node computer 111 and the port 121 sending out the information. This time moves in a clockwise direction in fig. 2. Since the network transmission time NWSZPKT 202 of the smart port 121 (and the transmission times of the other ports 122, 123 and of the other ports of the TT star coupler of the closed TT network connected via the connection line 100) is based on a time base of high accuracy (an accuracy of more than 1 μ sec is easily achievable with corresponding hardware support) [6], the information scheduler can, due to this accurate time base, enable scheduling of short-time and collision-free transmission between TT information and TT information in the closed TT network, enabling high rates of useful data in the TT network. After the network sending time NWSZPKT 202 is set, in the second stage, the scheduling of the time KNSZPKT 201 for the node computer to send TT information is set. Since KNSZPKT 201 is interpreted by the system time of the node computer, but network transmission time NWSZPKT 202 is interpreted by the system time of the TT star coupler, a predetermined degree of accuracy of the time synchronization between the node computer and the star coupler must be taken into account when scheduling node computer transmission time KNSZPKT 201 so that the start of the message arrives at the port of the TT star coupler on time before system transmission time NWSZPKT 202 even if close to the worst synchronization. The time gap between KNSZPKT and NWSZPKT must be greater than this time synchronization accuracy [6] plus twice the transmission duration of the first bit of the information and optionally the information preprocessing duration in the TT star coupler port. Furthermore, it must be ensured that the remainder of the information arrives at the port before it must be sent, so that the sending action is not interrupted once started.
If the node computer's system time is not synchronized with the system time of the TT star coupler, then KNSZPKT 201 is random. In this degenerate case, information arriving from a node computer is delayed in the port of the TT star coupler until the next periodic NWSZPKT 202 of such TT information is reached. If multiple TT messages of the same type arrive within a single message cycle, no further TT messages are issued and a fault message is generated at the diagnostic computer, since obviously a fault has already occurred.
Since it is assumed that all TT star couplers are provided with high-precision time synchronization supported by the respective hardware, the precision with which time synchronization can be achieved between the TT star couplers of the TT network and the node computers depends mainly on the type of time synchronization in the node computers. If time synchronization is achieved by software in the node computer, it may be difficult to obtain an accuracy of more than 50 musec. However, much higher accuracy can be achieved if the node computers are provided with dedicated synchronization hardware (i.e., time synchronization hardware corresponding to IEEE 1588 Standard [7 ]). The invention then supports different levels of end systems with different commercial ethernet controller offerings on the market without reducing the efficiency of the useful data of TT ethernet messaging in the TT network.
The characteristics of TT messages (e.g., KNSZPKT 201 and NWSZPKT 202, and optionally predicates of messages examined by the star coupler) must be known prior to sending TT messages, these characteristics can either be statically set by an offline scheduler before runtime, or dynamically determined online at the prompt of a node computer just before sending a message If the node computers support this synchronization standard, the time gap between KNSZPKT 201 and NWSZPKT 202 can be kept short on the basis of a high degree of accuracy in the system time of the node computers, thereby reducing the latency of real-time information transmission. Long transmission latencies must be accepted if this kind of hardware support on time synchronization is not given in the node computer.
If the TT Ethernet service node sends out the network sending time NWSZPKT 202 of all TT information to all TT star couplers affected by information transmission in advance, before the TT information arrives, the TT star coupler can keep a transmission channel of ET information required by [4] TT information transmission clear at the time of NWSZPKT in time, and therefore TT information can be transmitted on a reserved vacant line of the NWSZPKT with the least waiting time. This minimal information latency in the star coupler-only a few bits of delay (i.e., only a small amount of musec in a 100 mbit/sec ethernet system) is particularly important in systems containing multiple TT star couplers, such as by cascading many TT star couplers together to support bus wiring and simultaneous arrival of TT information in multiple node computers.
To ensure that after a temporary failure of the TT star coupler, all NWSZPKTs of global time, TT information, and all characteristics required to identify the failure of ethernet information expected to be received at the port, reappear within the intelligent ports 121, 122, 123 of the TT star coupler at a pre-specified restart time in the TT star coupler, the global time and TT information characteristics are periodically transmitted from one or more service nodes to the TT star coupler. The TT star coupler 101 periodically transfers the global time to the node computers 111, 112, 113 directly assigned to it so that the node computers can synchronize to the global time.
In a distributed real-time system, each node computer 111, 112 and 113 and the star coupler 101 form a fault-tolerant unit (FCU), i.e. they only show the immediate consequence of a fault (software or hardware) within the FCU result. The FCU may indirectly have a detrimental effect on the communication system and other node computers through fault information (in value or time range). The faulty FCU must be isolated. If only one of the node computers 111, 112, 113 fails, no assumptions can be made about the state of the failed node computer under normal circumstances. Thus, fault isolation can only be achieved when two independent FCUs are present, one FCU (e.g., node computer 111) showing a fault condition, while an independent second FCU (e.g., TT star coupler 101) recognizes the fault condition and prevents the fault from propagating. In accordance with the invention, a distinction is made between trusted (trusted) units and untrusted (untrusted) units. Assume that the star coupler 101 and TT ethernet service node are trusted, while the node computers 111, 112, 113 are typically untrusted. The trusted star coupler 101 prevents failures of the node computers 111, 112, 113 in hardware or software, whether accidental or intentional (security flaws) from adversely affecting the temporal behavior of the transmission of TT information by a non-failed node computer. To prevent a malicious sender from being able to send incorrect TT message characteristics to the TT star coupler, these message characteristics are transmitted from the TT service node to the TT star coupler in an encrypted encoded form.
In addition, a window of expected receipt with a start time 210 and an end time 211 may be predetermined in the information schedule so that faults within the time range of the mode computer 111 may be identified by the port 121 of the independent TT star coupler 101. if a Fault is identified by the port 121, the port 121 will send ET diagnostic information to the diagnostic computer. since the node computer 111 and the designated port 121 are disposed in two independent Fault Containment zones (Fault Containment regions), the situation where only one source of a fault causes a fault and the fault detection is switched off at the same time is eliminated. For this reason, observing the state of the node computer 111 at the designated intelligent port 121 is a particularly effective method of detecting failure [3 ].
The intelligent port assigned to the node computer (e.g., port 121 associated with node computer 111) may be cryptographically encoded before the arriving TT information is sent into the network. The port 121 must accordingly decode all encoded information arriving from the network before the information is transmitted to the node computer 111. The management of encoding and decoding of TT information is performed together with the encoded ET information.
Fig. 3 shows a possible structure of Time Triggered (TT) ethernet messages. The domains 301-305 and 310-312 are predetermined in the Ethernet standard [5 ]. Consistent with IEEE ethernet standard management, all protocol-specific TT ethernet messages contain bit patterns 88d7 within tag type field 305 to enable clear identification of each TT ethernet message worldwide. The temporal relationship describing the application information and the TT information may be chosen and thus may have an exactly defined time status. This information retains a predefined tag type field. By way of example, FIG. 3 provides a two-byte TT control field 306 in a possible, but not the only possible, embodiment of the present invention. The first byte of field 306 contains control information for the TT information, such as whether the TT information is synchronous or not. By means of another bit in the TT control field 306, a distinction is made between whether the information is periodic TT information or sporadic (sporadic) TT information. Periodic TT information is issued within each epoch assigned to such TT information. Occasional TT information is not sent out in every epoch assigned to such TT information. When the periodic TT information is interpreted by the receiving node as an important symbol as the transmitting node, its failure means a temporary or permanent failure at the transmitting node, which is not the case with the sporadic TT information.
The second byte of field 306 contains the length of TT ethernet information, which is in units of 8 bytes. The field 307 (period ID) is set at the position of the current period in the global time at a 16-bit level. KNSZPKT 201 interpreted by the node computer's system time is contained in field 307. The network transmission time NWSZPKT 202 interpreted by the system time of the TT star coupler is contained in field 308. The NWSZPKT also clearly defines the type of TT ethernet information, i.e., the information ID. As already mentioned above, the network transmission time NWSZPKT 202 for all TT information must be scheduled by the scheduler so that there is no conflict of TT information in the predetermined TT network. By contrast, it is possible that multiple node computers may send out their TT information at the same KNSZPKT 201 at the port of the TT star coupler assigned to them.
Finally, once again summarizing: the object of the present invention is to improve the efficiency and security of useful data when a distributed real-time computer system uses the commodity of an Ethernet controller by using TT Ethernet information communication through a plurality of node computers by one or more communication channels. To achieve this goal, a distinction is made between the node computer transmission time (KNSZPKT) and the network transmission time (NWSZPKT) of the information. The KNSZPKT interpreted by the system time of the sending node computer must be present in time before the NWSZPKT, so that in any event (i.e., even if the system times of the node computer and the TT star coupler are at the limits of the accuracy gap), the start of the message has reached the TT star coupler at the time of the NWSZPKT interpreted by the system time in the TT star coupler. It is proposed to tune the TT star coupler so that information arriving from the node computer is delayed in the intelligent port of the TT star coupler until the NWSZPKT, so that it can be sent into the TT network exactly at the NWSZPKT.
The above-described embodiments of the invention represent only one of many possibilities for implementing the invention.
Cited patents:
US 5,694,542 patent published at 12.12.1989: a loosely coupled distributed computer system with real-time precision node synchronization is provided.
EP 0658257 patent No. 12/18/1996: a communication control unit and method for transmitting information.
EP 1222542 patent No. 4/16/2003: an information distributor device with an integrated protection device.
EP 1512254 patent No. 5/10/2005: time Triggered (TT) ethernet.
Other references:
IEEE ethernet standard 802.3, URL: http:// standards. ie
Kopetz, H. (1997). Real-time systems and design principles of distributed embedded applications; ISBN: 0-7923-9894-7 Boston Kluwer Academic Publishers.
IEEE standard 1588 for clock synchronization, URL: http:// standards. ie
Claims (20)
1. A communication method for transmitting TT Ethernet messages in a distributed real-time computer system, the system comprising a plurality of node computers (111, 112, 113), wherein each node computer carries at least one Ethernet controller which is directly connected via data lines to ports (121, 122, 123) of a time-triggered star coupler (101) explicitly assigned to the node computer, and the plurality of time-triggered star couplers are directly or indirectly connectable to each other via one or more data lines (100) to form a closed time-triggered network,
characterized in that network transmission times (309) are pre-allocated to respective TT Ethernet messages by a scheduler, wherein a port (121) of the TT star coupler delays a TT Ethernet message arriving from the node computer (111) to a next network transmission time of such TT Ethernet messages in its system time, and the port (121) transmits the TT Ethernet message into the TT network at the network transmission time or transmits the TT Ethernet message into the TT network within a certain time period starting from the network transmission time upward.
2. A communication method according to claim 1, characterized in that a periodic node computer send time (308) is preassigned by the scheduler to a type of TT ethernet message, the node computer send time being interpreted by the system time of the sending node computer and setting a predetermined send time of the message from the node computer to a designated port of the TT star coupler, wherein the time difference between the node computer send time and the network send time must be greater than the accuracy of the time synchronization between the node computer's system time and the system time of the TT star coupler plus twice the duration of the delivery of the first bit of the message in the port of the autonomous TT star coupler and the optionally pre-processing duration of the TT ethernet message, so that the start of the message is ready in time for transmission into the TT network before the network send time, so that the entire message is constantly waiting for further continuous input into the TT network when appropriate.
3. A communication method according to claim 1, characterized in that the periodic network transmission time (308) is included in the received TT information.
4. A communication method according to claims 1-3, characterized in that the periodic node computer transmission time (309) is included in the TT information.
5. The communication method according to one of claims 1 to 4, characterized in that the intelligent port of the TT star coupler receives from the independent TT Ethernet service node the global time and the characteristics of the TT information expected at the port within a specified time range, and dynamically checks whether the arriving TT information corresponds to these pre-specified characteristics.
6. Communication method according to one of claims 1 to 5, characterized in that the intelligent port of the TT star coupler dynamically checks whether the characteristics expected to be within the specified value range of the port's TT information and the arriving TT information correspond to these pre-specified characteristics.
7. The communication method according to claim 5 or 6, characterized in that the intelligent port of the TT star coupler periodically receives from one or more independent TT Ethernet service nodes the global time and the specified characteristics of the TT information to be received, so that the TT star coupler again has the information of all the states required for fault detection to be reached within a period of time after a temporary fault of the TT star coupler.
8. Communication method according to one of the claims 5 to 7, characterized in that the independent TT Ethernet service node transmits the TT information properties in encrypted encoded form to the port of the TT star coupler.
9. The communication method according to one of claims 1 to 8, characterized in that the intelligent port of the TT star coupler sends ET ethernet information to the diagnostic computer if one or more specified characteristics of the arriving TT information violate a time range or a value range.
10. Communication method according to one of the claims 1 to 9, characterized in that the TT star coupler (101) periodically transmits the global time to the node computers (121, 122, 123) directly connected to the TT star coupler.
11. Communication method according to one of the claims 1 to 10, characterized in that the intelligent port of the TT star coupler cryptographically encodes the arriving TT information.
12. Communication method according to one of claims 1 to 11, characterized in that the intelligent port of the TT star coupler assigned directly to the node computer decodes the information arriving from the network before it is sent to the node computer.
13. The communication method according to claim 11 or 12, wherein information required for coding management of TT information coding or decoding is transmitted as encrypted ET information.
14. The communication method according to one of claims 1 to 13, characterized in that the TT star coupler duly clears the transmission channel required for the transmission of TT information between the TT star coupler and the node computer for the expected TT information on the basis of predicting the future network transmission time of the TT information.
15. Communication method according to one of the claims 1 to 14, characterized in that a distinction is made between periodic and sporadic (periodic) TT information.
16. A method of calculating the transmission time of TT information in a communication method according to claims 1-15, characterized in that in a first phase, the good accuracy of the system time of the TT star coupler is taken into account, a time-conflict free schedule of the network transmission times of all TT information is calculated, and in a second phase, the node computer transmission time of TT information to be transmitted by the node computers is determined on the basis of the time schedule, wherein the individual accuracy of each individual node computer and optional preprocessing of information assigned to this information in the port of the TT star coupler is taken into account.
TT ethernet service node computer, characterized in that it performs the method of claim 16 online and dynamically.
18. TT star coupler for propagating ethernet messages in a communication method according to one of the claims 1 to 15, characterized in that the TT star coupler comprises a central switch (102) and a plurality of intelligent ports (121, 122, 123), wherein each intelligent port (121, 122, 123) contains an autonomous intelligent subsystem with its own CPU and memory, and that these subsystems are additionally arranged to process arriving ethernet messages in parallel.
19. A TT star coupler for spreading ethernet messages according to claim 18, characterized in that the time systems of all TT star couplers constituting the TT network are synchronized in a highly accurate manner.
20. A TT star coupler for propagating TT ethernet messages, characterized in that the TT star coupler carries out one or more of the processing steps specified in claims 1-15.
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| ATA558/2007 | 2007-04-11 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| HK1143473A true HK1143473A (en) | 2010-12-31 |
Family
ID=
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101707954B (en) | Communication method and device for efficient and secure transmission of tt Ethernet messages | |
| US11157429B2 (en) | Time and event based message transmission | |
| US7839868B2 (en) | Communication method and system for the transmission of time-driven and event-driven Ethernet messages | |
| EP1490998B1 (en) | Method and circuit arrangement for the monitoring and management of data traffic in a communication system with several communication nodes | |
| CN102360206B (en) | Control system with a plurality of spatially distributed stations and method for transmitting data in said control system | |
| CN101632262B (en) | Node of a distributed communication system, node and monitoring device coupled to such communication system | |
| Ferreira et al. | Combining operational flexibility and dependability in FTT-CAN | |
| CN101346698A (en) | Bus monitor for users of communication systems and users of communication systems | |
| EP1629636B1 (en) | Time-triggered communication system and method for the synchronized start of a dual-channel network | |
| US10193705B2 (en) | Method, communication network, and control unit for the cyclical transmission of data | |
| US20160218946A1 (en) | Method for monitoring a first node in a communications network and monitoring system | |
| KR20040078113A (en) | Fault-tolerant clock synchronisation | |
| HK1143473A (en) | Communication method and device for efficient and secure transmission of tt ethernet messages | |
| Maier | Event-triggered communication on top of time-triggered architecture | |
| JP2005045772A (en) | Ensuring of maximum reaction times in complicated or distributed safe and or non-safe systems | |
| US7480832B2 (en) | Centralized error signaling and logging | |
| Ademaj et al. | Fault-tolerant time-triggered ethernet configuration with star topology | |
| Bertoluzzo et al. | Application protocols for safety-critical CAN-networked systems | |
| JPH1069448A (en) | Deterministic network protocol | |
| Bauer et al. | A hard real-time bus arbitration protocol based on EIA-709 | |
| He et al. | Time division multiple access based on CAN |