Multiplex communications – Fault recovery – Bypass an inoperative switch or inoperative element of a...
Reexamination Certificate
1998-08-21
2002-03-19
Hsu, Alpus H. (Department: 2662)
Multiplex communications
Fault recovery
Bypass an inoperative switch or inoperative element of a...
C370S227000, C370S397000, C709S239000
Reexamination Certificate
active
06359857
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to methods of triggering a rerouting of data in a packet based telecommunication system, to methods of bypassing faults in such systems, to such systems, nodes for such systems, and to software for such methods and for such nodes or systems.
BACKGROUND TO THE INVENTION
As explained in COM 13-R 7 ( March 1997) ANNEX 5 (to the report of WP 3/13, of ITU, “ATM Network Survivability Architectures and Mechanisms” network survivability can be divided into two broad categories, protection and restoration. Restoration includes reconfiguration, centrally controlled, and self healing, having distributed control, but not using completely dedicated bypass resources. The present invention is concerned with protection, which, for speed of operation has a distributed control architecture, and dedicated bypass paths.
Protection switching is concerned with minimising disruption to data traffic, at the expense of costly provision of dedicated paths with free bandwidth to enable data traffic to be switched instantly to the free path when necessary. One constraint which becomes more important, as bandwidth and data transmission reliability requirements increase, is the delay in detecting a need to switch, to trigger the protection, e.g. when a fault occurs.
Also, protection switching can occur at different layers in the network hierarchy. Coordination between layers may be necessary. Also, as connections are made over longer distances, delays in passing a trigger from a monitor at the destination, to the source, where the protection switching takes place, have meant that segmented protection switching has been used, for increased speed and efficient use of resources, particularly at lower layers of the network hierarchy.
Where multiple segments are monitored, it may be necessary for an alarm indication signal (AIS) to be sent to warn downstream monitors that a fault has been identified already, so the downstream monitors need not raise their own alarms as the consequences of the first fault propagate downstream.
Although the description hereinbelow will make use of ATM networks to show the, principles of the invention, they are clearly of broader applicability, e.g. to frame relay, or with appropriate modifications, to connectionless, networks, such as I.P.
Insertion of an AIS cell at the ATM layer, when a fault is detected, is shown in U.S. Pat. No. 5,461,607 (Miyagi et al). How a fault indication propagates up through network hierarchy layers from physical and transport layers up through the ATM layer to a data terminating equipment, is shown in U.S. Pat. No. 5,343,462 (Sekihata et al). Dealing with the alarm inpacket form rather than as lower level data, and doing the bypassing at the packet level makes it possible to arrange bypass paths at a more granular level on a connection basis, rather than having to bypass all the data on a link. This means the provisioning of these paths is more efficient and flexible, e.g. high priority connections can be separated, and allocated dedicated empty bypass paths, while lower priority connections or less sensitive connections might have to wait until other traffic is cleared from their bypass paths. Alarms causing the triggering of bypassing should be maintained at the packet level to avoid breaching the principle of passing alarms up the hierarchy, but never downwards where more data is multiplexed.
The problem of unnecessary protection switching occurring along an ATM VP in which multiple protected segments are provided, is described in contribution D47 of Q6/13, ITU meeting Turino, Italy, Jun. 16-20 1997. Examples in which an AIS cell is detected at more than one sink are shown. The sink is unable to determine from the AIS whether the fault is within its protection segment, or before it. If there is any dialogue of messages to determine the answer, or if a hold off is implemented, to wait and verify that the fault has not been bypassed already by a protection path in a preceding segment, before triggering protection, the delay would violate the requirement for fast protection switching.
One solution to the problem is shown in contribution D49 of Q6/13, ITU meeting Jun. 16-20 1997, Turino. On failure, an AIS cell is inserted and sent to the end of the connection. A bit in the cell is charged at the end of the segment in which failure occurred. The sink of the domain containing the failure triggers protection switching to bypass the failure. Other sinks further along the connection pass on the AIS cell, but know not to trigger their protection switching if they detect the charged bit in the AIS cell.
In contribution D48 to the above referenced ITU meeting nested protection schemes are catered for. The cell keeps a record of nesting level by recording how may sources or sinks it passes through, to enable the correct sink to trigger its protection as desired, whether that be the sink for the innermost of the nested protection schemes, or any other which covers the failed part.
It is known from contribution D50 of Q6/13 ITU meeting June 16-20 Turino, Italy to modify bits within the defect type indicator byte of the information field of the e-t-e AIS cells, to achieve recordal of status of the nesting level.
One problem with all these known arrangements is the requirement that each segment has the capability to extract, modify and reinsert the AIS cell. This results in greater complexity and cost, and may delay the throughput of data traffic, particularly if the sequence of the traffic is not disturbed.
In U.S. Pat. No. 5,212,475, a synchronous digital network is shown in which an alarm inhibit signal is generated at the physical layer when a fault is identified. The location of a fault is reported back to a central network management system. The inhibit signal is sent downstream to inhibit alarm generators downstream. This signal is modified by the insertion of a fault address message. Downstream signal distributors recognise the address and use it to determine whether to send a fault report message back to the central network management. system. They do not do so if the address is in the segment preceding a previous distributor. This means the alarm inhibit signal does not inhibit genuine alarms from other faults downstream. The central network management system triggers protection switching according to the messages it receives. However such an arrangement will not provide protection switching which is fast enough for many applications. Also, the requirement that every segment be able to extract and modify the inhibit signal, leads to increased complexity, particularly for high speed systems. The document does not refer to protection at a packet level.
SUMMARY OF THE INVENTION
It is an object of the present invention to provide improved methods and apparatus.
According to one aspect of the present invention there is provided a method of triggering a rerouting of data in a packet based telecommunication system, the system comprising a main data path and at least one bypass path, for bypassing a portion of the data path, the portion and the respective bypass defining a protection domain, the system comprising nodes for each of the domains, for monitoring respective domains and for issuing alarms in the form of packets, to other nodes downstream, with a domain identifier indicating the respective domain in which the alarm originated, the method comprising the steps of:
detecting at a given one of the nodes an alarm issued from a node upstream of the given node and a corresponding identifier;
determining at the given node whether to trigger a rerouting of the data along one of the bypass paths which bypasses the domain monitored by the given node, on the basis of the detected domain identifier.
An advantage of using the identifier is that downstream nodes can determine more easily whether the alarm is caused by a domain which has a bypass path triggered by another node upstream. Thus the problem of unnecessary triggering, described above, can be overcome without the considerable additional complexity, cost,
Ahmad Khalid
Martin David W.
Ye Yucheng
de Wilton Angela C.
Haszko Dennis R.
Hsu Alpus H.
Nortel Networks Limited
Qureshi Afsar M.
LandOfFree
Protection switching trigger generation does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Protection switching trigger generation, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Protection switching trigger generation will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2834007