Multiplex communications – Fault recovery – Bypass an inoperative channel
Reexamination Certificate
1999-06-07
2003-12-09
Nguyen, Steven H. D (Department: 2665)
Multiplex communications
Fault recovery
Bypass an inoperative channel
C370S218000, C370S236000, C370S410000, C709S237000
Reexamination Certificate
active
06661773
ABSTRACT:
TECHNICAL FIELD
The present invention relates to a data communication network, and more particularly, relates to a method for detection of stale cells following route changes in a data communication network.
BACKGROUND
A data communication network is generally consisted of a network of nodes connected by point-to-point links. Each link may be a bi-directional communication path between two connect nodes (stations) within the network. Data may be transmitted in groups called cells or packets from a source to a destination, often through intermediate nodes. In many data communication networks, cells (or data packets) between two endpoints (e.g., end stations such as computers, servers and/or I/O devices) may transverse the network along a given path to ensure that cells are delivered in the order transmitted and without duplication. A control mechanism may be used to permit the re-transmission of a cell or group of cells if the original cell is corrupted in the network due to network errors, including, for example, random noise.
If a particular link between two connect nodes fails within the infrastructure of the network, retransmission cells may be re-routed by the control mechanism via an alternate link that does not include the failed link (assuming such a link exists). After the failure is detected, a new route chosen by the control mechanism, duplicate old cells may still exist in the network that have not yet transversed the old route. These cells may be referred to as stale cells. These stale cells may interact with cell retransmission which can result in delivery of duplicate cells. The duplicate cells must be eliminated in order to prevent corruption of data communication between two endpoints in the network.
Currently, there may be two techniques commonly employed to address the duplicate or stale cell problem. The first technique seeks to include a sequence number in each cell at the transmitting endpoint. This sequence number must be sufficiently large to ensure that a unique number may be assigned to every cell that may exist in the network infrastructure at any given time. Using the sequence number, the receiving endpoint may discard any cell that contains a sequence number that has been previously-received. The major disadvantage of this technique is that, in large data communication networks, the overall performance of the network can be significantly degraded, e.g., significant computing time can be spent in bookkeeping choirs. Moreover, in many data communication networks that support multiple and/or simultaneous communication flows between two endpoints, the large sequence number can significantly affect the cost of implementing the context store required for each flow.
The second technique seeks to ensure a quiet period just prior to, and for a time period after, the network is configured with the new route. This quiet period must be sufficiently long to ensure that all cells that were in the network infrastructure at the time of the re-route have exited the network. During this quiet period, endpoints affected by the re-route may be prohibited from injecting new cells into the network. There may be two significant limitations to this technique, however. First, the endpoints of the network must be coordinated to prevent injection of new cells into the network. This required participation of the endpoints may increase the complexity of the switching algorithm. Second, the required quiet time may be unacceptably long, particularly in large data communication networks which operate at significant network speeds.
Therefore, there is a need for a more flexible, cost-effective and performance-efficient technique for recovery from faults within a network infrastructure which requires neither a quiet period, nor that the transmitting endpoint be aware that the route through the network infrastructure has changed for data communications.
SUMMARY
Accordingly, various embodiments of the present invention are directed to a data communication network and a method for detection of stale cells following route changes in a data communication network. The data communication network comprises a transmitter, which upon detection of a failure in a route of a network, retransmits data tagged as resent data along a different route; and a receiver, which upon detection of tagged data, utilizes tagged data for data communications while discarding previously transmitted data that are not tagged to avoid data duplication.
REFERENCES:
patent: 5222061 (1993-06-01), Doshi et al.
patent: 5398236 (1995-03-01), Hemmady et al.
patent: 5436886 (1995-07-01), McGill
patent: 5568476 (1996-10-01), Sherer et al.
patent: 5754754 (1998-05-01), Dudley et al.
patent: 5999515 (1999-12-01), Ohashi
patent: 6226298 (2001-05-01), Holden et al.
patent: 6272107 (2001-08-01), Rochberger et al.
patent: 6337852 (2002-01-01), Desnoyers et al.
patent: 6396835 (2002-05-01), Nagata
patent: 6404739 (2002-06-01), Gonno
patent: 6480897 (2002-11-01), Desnoyers et al.
patent: 6496481 (2002-12-01), Wu et al.
Cameron Donald F.
Pelissier Joseph E.
Blakely , Sokoloff, Taylor & Zafman LLP
Intel Corporation
Nguyen Steven H. D
Phan M.
LandOfFree
Method for detection of stale cells following route changes... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for detection of stale cells following route changes..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for detection of stale cells following route changes... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3108109