WO2004023302A2 - Verfahren und apparat zur fehlererkennung in einem verteilten echtzeitcomputersystem - Google Patents
Verfahren und apparat zur fehlererkennung in einem verteilten echtzeitcomputersystem Download PDFInfo
- Publication number
- WO2004023302A2 WO2004023302A2 PCT/AT2003/000256 AT0300256W WO2004023302A2 WO 2004023302 A2 WO2004023302 A2 WO 2004023302A2 AT 0300256 W AT0300256 W AT 0300256W WO 2004023302 A2 WO2004023302 A2 WO 2004023302A2
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- node
- message
- function group
- bit
- sending
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
- G06F11/0703—Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation
- G06F11/0751—Error or fault detection not based on redundancy
- G06F11/0763—Error or fault detection not based on redundancy by bit configuration check, e.g. of formats or tags
Definitions
- the invention relates to a method for consistently detecting errors in a distributed real-time computer system, comprising a plurality of node computers, e.g. the five node computers, each having at least one grain communication controller and one host computer each, exchanging messages via a real-time, turn-based communication system consisting of one or more communication channels, transporting one or more redundant physical messages for each logical message and where all the nodes have access to a globally synchronized time.
- node computers e.g. the five node computers, each having at least one grain communication controller and one host computer each, exchanging messages via a real-time, turn-based communication system consisting of one or more communication channels, transporting one or more redundant physical messages for each logical message and where all the nodes have access to a globally synchronized time.
- the invention relates to a communication controller of a computer node of a distributed computer system.
- a distributed real-time computer system consists of a number of node computers, which exchange messages via a real-time communication system and which observe and influence a technical process via local interfaces.
- Such distributed real-time computer systems are increasingly being used to control safety-critical applications. Examples of such applications are the control of the brakes in a motor vehicle or a flight control system in an aircraft. In these applications, it is of great importance that a occurring error, which can lead to the loss of data consistency in the distributed system, is quickly detected. Examples of such errors are the loss of messages, the failure of a node computer, or an error in the process peripheral assigned to the node computer.
- a particularly common and important error class in distributed real-time systems is the class of crash / emissions (CO) errors.
- the error code of the CO errors includes the loss of messages or the total failure of node computers.
- Many of the known grain communication protocols contain mechanisms to detect CO defects [6].
- a method that can detect CO errors in a distributed real-time system has been disclosed in patents [1-4]. This method assumes that all computers of a subsystem (cluster) are involved in all functions of the system and that the inconsistency caused by a CO error of an account computer is remedied by the fact that the affected node computer sends no further messages until it reintegrated successfully again.
- This type of error detection and error handling is considered to be too restrictive in some applications, as well as the failure of a single function of a node computer, which is involved in several functions, leads to the failure of the entire node computer.
- a node computer can be a member of several functional groups and where each functional group is assigned a dedicated functional group member field whose bit length corresponds at least to the number of node computers in this function group, and where each node in a function group is uniquely assigned a specified bit of the function group member field, and where each node computer has a local version of the function group member field of each function group in which it participates, and, where before sending a message, the sending node sets the bit associated with it in its local field member field to TRUE if the sending node is functional in that foot and, where before sending a message, the sending node sets its associated bit in its local functional group member field to FALSE if the sending node considers its functionality in that function group to be non-existent, and where the sending node collects all of its local functional group member fields copied into a message to be sent to transmit to the recipients.
- the fault detection is performed functionally, wherein a node computer can be involved in several functions.
- Data consistency is ensured within function groups.
- the four node computers that control the brakes on the four wheels may form such a functional group.
- a Radknotenrechner can also be involved in another function that has nothing to do with the immediate braking function.
- An error in this "other" function should not lead to an error of the brake function From the point of view of the application, it is important that within a function group a consistency loss is detected quickly, the operation of the other independent function groups, in which a node computer also involved is, but not affected.
- the receiver of a message determines the bit associated with a transmitting node in each of its local functional group member fields immediately after the expected time of receiving a message as follows: (1) If none of the sending node's redundant messages has arrived correctly syntactically (2) If the bit of a function group member field relating to the message sender in the received message is set to FALSE, then the receiver sets the corresponding bit in its corresponding one (3) If the sum of the transmission time of the message and the known transport latency deviates significantly from the reception time observed by the receiver, then the receiver sets the bits allocated to the transmitter in all its corresponding local function group membership members to FALSE. In all other cases, the receiver sets TRUE to the bits corresponding to the sender in all its local function group member fields.
- each member of a function group manages a functional group related consistency variable, which is set to TRUE immediately after sending and is set to the value FALSE, as soon as contained in a received message function group member field of a successor of the sending node Content of the corresponding local function group member field.
- each group of a function group manages a function group-related acknowledgment variable which is set to the value FALSE immediately after transmission and which is set to the value TRUE as soon as the bit of the function group contained in a received message relates to the originally transmitting node member field in a received message of a successor of the originating node is set to TRUE.
- the node computers participating in a function group are informed about one or more configuration messages, which node computers form a function group, and when which messages are to be expected from which node computers.
- the configuration messages are encrypted.
- the periodicity of the transmission within a function group is specified by the specification of a periodicity bit in the time format presented in the introduction and the period starts at the time at which all bits of the time value to the right of the periodicity bit are not set.
- the periodic transmission time of a message is determined by the specification of the offset at the beginning of the period.
- the communication controller of the sending node of a message sets its associated bit in all local function group member fields of the sending node immediately before sending to FALSE, if within the last Kir.unikationsêt that assigned to the sending node Bit of the feature group member field was not written by the host computer.
- the object according to the invention is furthermore achieved by a communication controller, mentioned in the introduction, of a computer node of a distributed computer system, in which according to the invention one or more of the above-mentioned method steps of the communication controller are implemented.
- the communication controller signals to the host computer via an interrupt line when a bit of a local function group member field changes to the state FALSE. It may also be advantageous if the communication controller signals the host computer via an interrupt line immediately before sending a message that the acknowledgment variable according to claim 5 has the value FALSE.
- the communication controller signals to the host computer via an interrupt line immediately before sending a message that the consistency variable according to claim 4 has the value FALSE.
- the invention disclosed herein provides the following advantages over the prior art distributed error detection technique, e.g. implemented in the TTP / C protocol [1-4]:
- a differentiated membership is introduced that allows nodes to continue to operate that are inconsistent in subfunctions.
- the membership algorithm is solved from the transmission protocol.
- the reliability of a given function is increased by the fact that the failure of another independent function in a node computer has no influence on the given function.
- the separation of the error detection function from the immediate protocol function allows the construction of fault tolerant systems with standard protocols, such as, for example ETHERNET, as the fault detection protocol can be implemented in a protocol layer over the standard data communication protocol.
- standard protocols such as, for example ETHERNET
- FIG. 1 shows the structure of a distributed computer system.
- a real-time communication system 109 is a communication system that transmits messages within a known real-time interval, the transport latency.
- the communication system may apply resource redundancy or time redundancy, ie, for each logical message provided by the communication controller 121, multiple redundant instances of the logical message, ie, multiple physical ones, may be simultaneously on multiple parallel communication channels or consecutively on a communication channel Messages to be sent. It is assumed that the recipient has been informed a priori of the transmission time of a message. This announcement of the transmission times can be done dynamically immediately before sending. The recipient may choose from the time of transmission and the known transport of the communication system, when the last correct instance of a message has to arrive at the latest at the receiver. If no message has arrived from the sender until then, then an error has occurred from the recipient's point of view (message loss).
- resource redundancy or time redundancy ie, for each logical message provided by the communication controller 121, multiple redundant instances of the logical message, ie, multiple physical ones, may be simultaneously on multiple parallel communication channels or consecutively on a communication channel Messages to be sent. It is assumed that the recipient has been informed
- the five node computers 111, 112, 113, 114 and 115 can perform several functions and therefore belong to different functional groups.
- the four node computers 111, 112, 113 and 115 can form the function group 1
- the four node computers 111, 112, 113 and 114 form the function group-2.
- the function of distributed clock synchronization [6] and function group-2 realize the function of a four-brake distributed braking system in a motor vehicle controlled by the node computers 111, 112, 113, and 114. If an error occurs in a brake, the clock synchronization function running in parallel in the computer node is not affected by this error unrnittelbaj.
- each of these functional groups is assigned its own functional group member field (FIG. 2) whose bit length corresponds at least to the number of node computers in the function group, and where each node computer involved in a function group has a particular bit, e.g. 201, 202, 203, 205, the function group member field is uniquely assigned.
- the function group member field is uniquely assigned.
- the value TRUE of the bit 201 that the function-1 in the node computer 111 is executed correctly and the value FALSE that the function-1 in the node computer 101 is not executed correctly.
- Each node computer 111 manages a local version of the constituent element of each functional group in which it participates.
- the round period (field 301 in FIG. 3) can be derived directly from the time available globally.
- Fig. 4 shows a time representation [5] standardized by the Object Management Group (OMG) consisting of 8 bytes.
- This time representation 401 uses the binary format, with the full physical second 402 representing the central reference. With this time format 401, any time within the next ten thousand years can be represented with an accuracy of approximately 60 nanoseconds.
- Each bit of this time format can be assigned a periodicity, which results from the position of the bit in the time format.
- the periodicity bit 403 is the 6th bit to the right of the full seconds bit 402.
- the periodicity bit 403 thus determines a rounding frequency of 1/26 seconds, which corresponds to 64 Hz.
- the beginning of the period is defined as the time at which all Bits to the right of periodicity bit 403 are not set in time format 401.
- the transmission time of a node computer can be precisely specified by the specification of the offset 410 from the start of the period. If the periodicity of each round equals a power of two of the smallest round, then it is possible to easily build large synchronous systems that have a multiplicity of clusters. In the different clusters different Kornrnunikationsprotokolle can be used.
- the sending account calculator (sender) 111 sets the bit associated with it in its local action group member field 201 to TRUE if the sender considers its functionality in this function group to have been established. Prior to sending a message, the sender 111 sets the bit associated with it in its local functional group member field 201 to FALSE if the sender considers its functionality in this function group to be non-existent. In the previous example with the defective brake, the transmitter will set its bit in function group field 1, the clock sync to TRUE and its bit in the function group member field, the brake system, to FALSE if the local brake is faulty.
- the sender Immediately before sending a message, the sender will copy all of its local member group member names into the message to be sent and send it to the recipients with the message.
- the receiver of a message sets the bit associated with the sender in each of its local function group member fields immediately after the expected time of receipt of a message from a sender, as follows: (1) If, by that time, none of the sender's redundant messages have arrived syntactically correct, then the receiver sets the value of the bit associated with the sender to FALSE in all its affected local function group member fields. A message is syntactically correct if the CRC check of the message does not indicate an error.
- the receiver sets the value of the bit associated with the transmitter in all its local function group member fields to FALSE.
- the observed time of reception significantly deviates from the expected time of reception if the difference between these two times, which must lie in the active intervals of a sparse timebase (see [6], p. 56), is greater than the duration of an active time interval of the sparse timebase ,
- the receiver sets TRUE to the corresponding bits in its local function group member fields.
- Each member of a function group can maintain a function group-related consistency variable that is set to TRUE immediately after transmission, and is set to FALSE as soon as the function group member field of a successor of the sender contained in a received message deviates from the content of the corresponding local function group member field. This rule detects an inconsistency in the system that could be observed within the last round.
- Each member of a functional group may manage a functional group-related acknowledgment variable which is FALSE immediately after transmission and which is set to the value TRUE as soon as the transmitter-related bit of the group membership field contained in a received message is included in a message received from a successor of the processor Transmitter is set to TRUE. This rule checks if any recipient has received the sender's message.
- the communication controller 121 of the sender of a message sets the bit assigned to the sender in all local member groups of the sender immediately before transmission to FALSE, in case that the sender-assigned bit of the function group member field was not written by the host computer 131 within the last communication round, then For example, in each round in each functional group member field, the host computer 131 is forced to set the bit associated with the transmitter to TRUE again, as long as that function is still met. This rule prevents a fail-silent failure of a function of the host computer 131 from being observed by the communication partners within a function group.
- two different function groups with corresponding function group member fields may be introduced.
- One function group monitors the function of the communication, while the other function group monitors the function of the application in the host.
- the disclosed method for consistent error detection in a distributed computer system may also be implemented in the hardware of the communication controller 121.
- it may be appropriate to signal the host computer 131 via an interrupt line 151 (error interrupt) when a bit of a local function group member field of the communication controller 121 changes from the state TRUE to the state FALSE. If no error interrupt occurs, the host computer 131 may assume that the communication within the functional groups of the host computer is functioning properly.
- the communication controller 121 may also signal the host processor 131 via an interrupt line 151 if, in the previous round, none of the correct recipients have acknowledged the last message sent by this node computer or if an inconsistency has been observed in the last round.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer And Data Communications (AREA)
- Debugging And Monitoring (AREA)
- Multi Processors (AREA)
- Hardware Redundancy (AREA)
Abstract
Description
Claims
Priority Applications (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| EP03793461A EP1535156B1 (de) | 2002-09-05 | 2003-09-03 | Verfahren und apparat zur fehlererkennung in einem verteilten echtzeitcomputersystem |
| DE50303229T DE50303229D1 (de) | 2002-09-05 | 2003-09-03 | Verfahren und apparat zur fehlererkennung in einem verteilten echtzeitcomputersystem |
| AU2003264154A AU2003264154A1 (en) | 2002-09-05 | 2003-09-03 | Method and device for recognizing errors in a distributed real-time computer system |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| ATA1332/2002 | 2002-09-05 | ||
| AT0133202A AT500050A2 (de) | 2002-09-05 | 2002-09-05 | Verfahren und apparat zur fehlererkennung in einem verteilten echtzeitcomputersystem |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| WO2004023302A2 true WO2004023302A2 (de) | 2004-03-18 |
| WO2004023302A3 WO2004023302A3 (de) | 2004-10-28 |
Family
ID=31953362
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/AT2003/000256 Ceased WO2004023302A2 (de) | 2002-09-05 | 2003-09-03 | Verfahren und apparat zur fehlererkennung in einem verteilten echtzeitcomputersystem |
Country Status (5)
| Country | Link |
|---|---|
| EP (1) | EP1535156B1 (de) |
| AT (2) | AT500050A2 (de) |
| AU (1) | AU2003264154A1 (de) |
| DE (1) | DE50303229D1 (de) |
| WO (1) | WO2004023302A2 (de) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2008151339A3 (de) * | 2007-06-11 | 2009-03-12 | Fts Computertechnik Gmbh | Verfahren und architektur zur sicherung von echtzeitdaten |
| US7715408B2 (en) | 2004-09-15 | 2010-05-11 | Tttech Computertechnik Ag | Method for establishing communication plans for a divided real-time computer system |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO1994006080A1 (de) * | 1992-09-04 | 1994-03-17 | Fault Tolerant Systems | Kommunikationskontrolleinheit und verfahren zur übermittlung von nachrichten |
| US5694542A (en) * | 1995-11-24 | 1997-12-02 | Fault Tolerant Systems Fts-Computertechnik Ges.M.B. | Time-triggered communication control unit and communication method |
| US5799146A (en) * | 1996-04-30 | 1998-08-25 | International Business Machines Corporation | Communications system involving groups of processors of a distributed computing environment |
-
2002
- 2002-09-05 AT AT0133202A patent/AT500050A2/de not_active Application Discontinuation
-
2003
- 2003-09-03 EP EP03793461A patent/EP1535156B1/de not_active Expired - Lifetime
- 2003-09-03 DE DE50303229T patent/DE50303229D1/de not_active Expired - Lifetime
- 2003-09-03 WO PCT/AT2003/000256 patent/WO2004023302A2/de not_active Ceased
- 2003-09-03 AT AT03793461T patent/ATE325382T1/de not_active IP Right Cessation
- 2003-09-03 AU AU2003264154A patent/AU2003264154A1/en not_active Abandoned
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7715408B2 (en) | 2004-09-15 | 2010-05-11 | Tttech Computertechnik Ag | Method for establishing communication plans for a divided real-time computer system |
| WO2008151339A3 (de) * | 2007-06-11 | 2009-03-12 | Fts Computertechnik Gmbh | Verfahren und architektur zur sicherung von echtzeitdaten |
| US8464065B2 (en) | 2007-06-11 | 2013-06-11 | Fts Computertechnik Gmbh | Procedure and architecture for the protection of real time data |
Also Published As
| Publication number | Publication date |
|---|---|
| EP1535156A2 (de) | 2005-06-01 |
| DE50303229D1 (de) | 2006-06-08 |
| ATE325382T1 (de) | 2006-06-15 |
| AU2003264154A1 (en) | 2004-03-29 |
| AT500050A2 (de) | 2005-10-15 |
| EP1535156B1 (de) | 2006-05-03 |
| WO2004023302A3 (de) | 2004-10-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP0732654B1 (de) | Verfahren zur fehlertoleranten Kommunikation unter hohen Echtzeitbedingungen | |
| EP2556633B1 (de) | Verfahren und apparat zur fehlertoleranten zeitgesteuerten echtzeitkommunikation | |
| DE3486148T2 (de) | Fehlertolerantes Übertragungssteuersystem. | |
| DE3788551T2 (de) | Mehrpunkt-Datenübertragung mit Überwachungssystem zum Erkennen von gestörten Aussenstellen. | |
| DE3685609T2 (de) | Kommunikationsverfahren fuer multiprozessor. | |
| WO2001013230A1 (de) | Verfahren zum erzwingen der fail-silent eigenschaft in einem verteilten computersystem und verteilereinheit eines solchen systems | |
| DE69921882T2 (de) | Verfahren zur Entdeckung und Lösung von Datenkorruption in einem UART-basierten Kommunikationsnetzwerk | |
| DE69802535T2 (de) | Aktive fehlererkennung | |
| EP2936747A1 (de) | Datenübertragung unter nutzung eines protokollausnahmezustands | |
| DE10316649A1 (de) | Sicherheitskommunikation auf einer einzigen Rückebene | |
| AT410490B (de) | Verfahren zur tolerierung von ''slightly-off- specification'' fehlern in einem verteilten fehlertoleranten echtzeitcomputersystem | |
| WO2002001305A1 (de) | Redundantes steuerungssystem sowie steuerrechner und peripherieeinheit für ein derartiges steuerungssystem | |
| EP3110061B1 (de) | Verteiltes echtzeitcomputersystem sowie verfahren zur erzwingung des fail-silent-verhaltens eines verteilten echtzeitcomputersystems | |
| EP0720337B1 (de) | Verfahren zur hochzuverlässigen und konsistenten Nachrichtenübertragung | |
| EP1535156B1 (de) | Verfahren und apparat zur fehlererkennung in einem verteilten echtzeitcomputersystem | |
| EP1064590A1 (de) | Verkürztes datentelegramm eines automatisierungssystems | |
| EP0965924B1 (de) | Anordung und Verfahren zum Übertragen von Adress-, Befehls- und/oder Datentelegrammen | |
| EP1399818B1 (de) | Verfahren und vorrichtung zur kommunikation in einem fehlertoleranten verteilten computersystem | |
| DE10161295A1 (de) | Verfahren zum Vermeiden eines Totalausfalls einer Verarbeitungseinheit für Protokollinformationen | |
| EP1686732A1 (de) | Verfahren und System zur Übertragung von Telegrammen | |
| EP1619849A1 (de) | Verfahren zum Synchronisieren eines verteilten Systems | |
| DE102021127310A1 (de) | System und Verfahren zur Datenübertragung | |
| DE102009016972B4 (de) | Kommunikationssystem zum dezentralen und autarken Überwachen und Steuern eines unterlagerten Bussystems | |
| DE3427350C2 (de) | ||
| EP2741453B1 (de) | Verfahren zum betreiben eines busgeräts einer gebäudeautomatisierungseinrichtung, sowie entsprechendes konfigurationsgerät und entsprechendes computerprogrammprodukt |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| AK | Designated states |
Kind code of ref document: A2 Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NI NO NZ OM PG PH PL PT RO RU SC SD SE SG SK SL SY TJ TM TN TR TT TZ UA UG US UZ VC VN YU ZA ZM ZW |
|
| AL | Designated countries for regional patents |
Kind code of ref document: A2 Designated state(s): GH GM KE LS MW MZ SD SL SZ TZ UG ZM ZW AM AZ BY KG KZ MD RU TJ TM AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IT LU MC NL PT RO SE SI SK TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG |
|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
| DFPE | Request for preliminary examination filed prior to expiration of 19th month from priority date (pct application filed before 20040101) | ||
| WWE | Wipo information: entry into national phase |
Ref document number: 92232003 Country of ref document: AT |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 2003793461 Country of ref document: EP |
|
| WWP | Wipo information: published in national office |
Ref document number: 2003793461 Country of ref document: EP |
|
| WWG | Wipo information: grant in national office |
Ref document number: 2003793461 Country of ref document: EP |
|
| NENP | Non-entry into the national phase |
Ref country code: JP |
|
| WWW | Wipo information: withdrawn in national office |
Country of ref document: JP |