Electrical computers and digital processing systems: multicomput – Computer conferencing – Demand based messaging
Reexamination Certificate
1998-07-21
2002-08-20
Barot, Bharat (Department: 2154)
Electrical computers and digital processing systems: multicomput
Computer conferencing
Demand based messaging
C709S201000, C709S202000, C709S205000, C709S220000, C707S793000, C707S793000, C707S793000, C707S793000
Reexamination Certificate
active
06438582
ABSTRACT:
FIELD OF THE INVENTION
This invention relates to parallel or distributed data processing systems and more particularly to a mechanism which avoids duplication of messages when processing transaction commit and rollback procedures.
BACKGROUND OF THE INVENTION
In a multi-node environment, a transaction which has been started on an originating, or coordinating, node will be sent to associated remote nodes, referred to as participant nodes. After the transaction has started at the coordinating node, C, a remote node, R, will be added to the coordinating node's participant node list when a request is sent to node R on behalf of the transaction. Standard procedure on a coordinating node is to maintain one participant list, which is continually updated as needed. Since more than one transaction may be executing in sequence, and less than all of the nodes on the system may be participating in each transaction, communications regarding a given transaction which are sent to all nodes on the participant list may arrive at nodes which are not participating in the transaction. Furthermore, on behalf of the same transaction, the remote participant node R may send a request on to another remote node, N. Node N is then added to a transaction participant list at node R. It is possible, however, that the coordinating node C does not have any awareness of node N. Therefore, should node C need to communicate information, such as a transaction rollback request, to all nodes which have participated in the transaction, node C will only have sufficient information for contacting those nodes in the node C participant node list. In the prior art, node C would provide the rollback request to node R, and node R would then be responsible for passing the rollback request on to any nodes, N, which are on its participant list for the transaction, as illustrated in FIG.
1
.
FIG. 1
shows that any given “subordinate” node, such as node
4
which has transaction ties to nodes
1
,
2
and
5
, could receive a given message generated at the coordinating node three times (one each from nodes
1
,
2
and
5
).
In the alternative, participant node R could provide its participant list to coordinating node C in response to a message, with node C then resending the message to all participant nodes. In such as scenario, each of nodes
1
,
2
and
5
, could receive the message two times, while node
4
could receive the message six times. The sending of redundant messages results in unnecessary bandwidth consumption and unnecessary duplication of message processing at both the coordinating node and the remote participant nodes.
It is therefore an objective of the present invention to provide a mechanism by which, in parallel or distributed processing systems, transaction information can be most efficiently communicated to relevant nodes.
SUMMARY OF THE INVENTION
This and other objectives are realized by the present invention wherein participant node lists are maintained for each individual transaction. In addition, local nodes maintain transaction participant node lists and provide same to the transaction coordinating nodes under certain circumstances. The coordinating nodes merge transaction participant node lists and selectively utilize the lists to minimize message duplication. Further, connection node lists for each application are maintained at coordinating nodes and are utilized for application rollback procedures.
REFERENCES:
patent: 5202971 (1993-04-01), Henson et al.
patent: 5291594 (1994-03-01), Sekiguchi et al.
patent: 5390302 (1995-02-01), Johnson et al.
patent: 5546582 (1996-08-01), Brockmeyer et al.
patent: 5557748 (1996-09-01), Norris
patent: 5799305 (1998-08-01), Bortvedt et al.
patent: 5872969 (1999-02-01), Copeland et al.
patent: 5884327 (1999-03-01), Cotner et al.
patent: 5958004 (1999-09-01), Helland et al.
patent: 5999712 (1999-12-01), Moiin et al.
patent: 6085295 (2000-07-01), Ekanadham et al.
patent: 6247023 (2001-06-01), Haiao et al.
Chang Amy
Hsiao Hui-I
Barot Bharat
Dougherty Anne Vachon
Ellenbogen Wayne L.
LandOfFree
Method and system for efficiently coordinating commit... 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 and system for efficiently coordinating commit..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for efficiently coordinating commit... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2921987