Multiplex communications – Fault recovery
Reexamination Certificate
1999-06-03
2004-04-27
Vanderpuye, Kenneth (Department: 2732)
Multiplex communications
Fault recovery
C370S242000
Reexamination Certificate
active
06728204
ABSTRACT:
BACKGROUND
The present invention relates to a distributing arrangement for distribution of messages in a communication system supporting communication of packet data. The invention also relates to a communication system including a number of distributing nodes each of which includes a number of processing means. The invention also relates to a method of fault tolerant communication of messages in a communication system supporting communication of packet data. Particularly the invention relates to an arrangement and a system respectively for fault tolerant communication of messages.
In data/telecommunication systems messages are sent between different processing means. It has become common to implement LAN (Local Area Network) technology within data- and telecommunication equipment. Two reasons therefore are that the LAN technology has become attractive since it is cost effective and that the LAN technology furthermore provides a good performance. However, to provide a satisfactory operation of the systems, the communication needs to be as fault tolerant as possible to introduce reliability, redundancy and freedom from single point failures. A single LAN segment is not sufficiently reliable since the central equipment is not redundantly implemented. In known systems so called FDDI/DAS (Fiber Distributed Data Interface/Dual Attach Station) solutions or alternatively high availability ATM (Asynchronous Transfer Mode) solutions are therefore used.
Both FDDI/DAS and ATM solutions are however expensive.
An implementation of a FDDI/DAS ring is for example known. In that case redundancy is provided for through connecting each processing means to the dual ring and the handling of faults is taken care of in the points of connection.
FDDI/DAS technology is expensive and furthermore it is limited to 100 Mbps making topology restraints on nodes which require a higher bandwidth necessary. Thus, in that case the costs are still further increased since more network interfaces are needed. Furthermore the degree of freedom in software configuration is somewhat limited.
Alternatively a high availability ATM (Asynchronous Transfer Mode) switch with built-in reducancy may be used. Such a solution is however even more expensive than the FDDI/DAS solution. In some technologies the fault tolerancy is handled by protocols on a high level or on the routing level. Furthermore, through the implementation of bridges between the different networks or network segments, fault tolerant communication is not provided everywhere and the communication with some processing means may be lost if a bridge is broken.
SUMMARY
What is needed is therefore an arrangement through which fault tolerant communication can be offered in a data/telecommunication system supporting communication of packet data. Still further an arrangement is needed through which fault tolerant communication can be provided and through which the amount of messages sent within the network, i.e. the network overhead produced hereby, can be kept low. Particularly an arrangement is needed through which it is possible to detect errors in equipment as well as on communication paths and through which it is possible to detect when a communication path etc. functions satisfactorily again and through which messages can be sent from one processing means to another with a high degree of availability, i.e. with a high probability of success. Still further an arrangement is needed through which faults can be detected in a fast and efficient manner thus enabling restoration of errors.
Particularly an arrangement is needed through which the provision of fault tolerance is transparent for high level protocols and on the routing level. Still further an arrangement is needed which allows the use of standard components (hardware). Particularly an arrangement is needed which is cheap and which has a high performance and which is reliable, shows a high degree of redundancy and which provides freedom from single point failures. Still further an arrangement is needed which enables an automatic error detection and failover as well as automatic recovery. Particularly an arrangement is needed which enables operator notification of failures particularly while the operation still is substantially unaffected.
Particularly a distributing arrangement for distribution of messages in a communication system is needed which meets the above mentioned requirements. Still further a communication system supporting communication of packet data is needed through which the above mentioned objects are achieved.
Further yet a method providing fault tolerant communication in a communication system supporting communication of packet data is needed through which the above mentioned objects are achieved.
Therefore a distributing arrangement for distribution of messages in a communication system is provided which comprises a number of processing means which communicate with each other through sending of messages. Each processing means is connected to all the other processing means via at least two communication networks or network segments which are physically separated from each other. For each processing means information holding means are provided for keeping information about all communication paths to all other processing means and this information is obtainable by way of sending first messages, particularly with a given frequency, from each processing means to all the other processing means using all of said at least two communication networks. Such first messages may be sent separately but they may also be piggy-backed on e.g. ordinary traffic messages, or a combination of both. Advantage may also be taken of information indirectly obtainable through the ordinary traffic that flows.
In an exemplary embodiment a predetermined time interval is given for each processing means and first information messages sent from one processing means substantially simultaneously to another processing means using at least two physically separated communication paths including different networks should arrive within the predetermined time interval. If they do so, the at least two different communication paths using different communication networks both function. In an advantageous embodiment the first information messages are special purpose broadcast messages and preferably each processing means send their respective special purpose broadcast messages on all networks or network segments substantially simultaneously. Alternatively the first messages consist of a burst of unicast messages. Particularly each processing means comprises means for detecting anomalies or asymmetries in the reception of such first messages from one and the same sending processing means and when such an asymmetry or anomaly is detected, the receiving processing means enters a decision process to localize the error and or to establish the kind of the error. In a particular embodiment means are provided for comparing the reception of first messages during a first time interval and the reception of first messages during a subsequent observation time intervals. If an anomaly or an asymmetry in the reception during the first observation time interval still remains in the subsequent observation time interval, the receiving processing means sends a second message, e.g. a unicast message containing information about an error or about the result of the sending of the first messages, to the sending processing means on all communication networks
etwork segments. In an alternative embodiment a second message, e.g. a unicast message, is sent to the sending processing means on all networks
etwork segments as soon as an error is detected. Still further, more than two observation periods during which one and the same anomaly remains could be the requirement for sending out a second, e.g. a unicast message containing an error announcement. With an asymmetry or anomaly is meant that not all first messages from one and the same sending processing means arrive within the time window of the receiving processing means. It is supposed that each process
Helander Lars-Erik
Pettersson Sten
Telefonaktiebolaget LM Ericsson
Vanderpuye Kenneth
LandOfFree
Arrangement, system and method relating to... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Arrangement, system and method relating to..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Arrangement, system and method relating to... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3219466