Summary-based routing for content-based event distribution networks

Abstract

A system arid method for enabling highly scalable multi-node event distribution networks through the use of summary-based routing, particularly event distribution networks using a content-based publish / subscribe model to distribute information. By allowing event routers to use imprecise summaries of the subscriptions hosted by matcher nodes, an event router can eliminate itself as a bottleneck thus improving overall event distribution network throughput even though the use of imprecise summaries results in some false positive event traffic. False positive event traffic is reduced by using a filter set partitioning that provides for good subscription set locality at each matcher node, while at the same time avoiding overloading any one matcher node. Good subscription set locality is maintained by routing new subscriptions to a matcher node with a subscription summary that best covers the new subscription. Where event space partitioning is desirable, an over-partitioning scheme is described that enables load balancing without repartitioning.

Description

FIELD OF THE INVENTION [0001] This invention pertains generally to computer networks, and, more particularly, to computer networks that use a publish / subscribe model to distribute information. BACKGROUND OF THE INVENTION [0002] Today's computer data networks span the globe and provide an ever increasing variety of information and types of information. A popular model for retrieving information is the request-response model. This is a model used, for example, by the World Wide Web: a Web client requests a Web page from a Web server and then waits until the Web server responds. This model is adequate for basic access to information, but as information consumers become more sophisticated, it quickly becomes inefficient for information consumers or information providers or both. As a general example, under the request-response model, a consumer only interested in changes to an item of information (e.g., a stock price) may be required to request the information over and over again until ...

AI Technical Summary

Benefits of technology

[0012] An event router node of an event distribution network maintains an imprecise summary of the set of subscriptions hosted by each matcher node. If the event router node is overloaded, it reduces the precision of the imprecise summaries. Reducing the precision of the imprecise summaries allows the event router to process each event faster. If the event router load falls beneath some high threshold, then it increases the precision of the imprecise summaries. Increasing the precision of the imprecise summaries reduces the amount of false positive traffic routed to a matcher node. False positive traffic makes a matcher node work harder. There is a balance point at some level of imprecision that optimizes the throughput of the event distribution network as a whole.

[0013] Subscriptions to be hosted by an event distribution network are divided among the matcher nodes of the event distribution network so as to provide good subscription locality to the set of subscriptions hosted by each matcher node, while at the same time avoiding overloading any one matcher node (i.e., ensuring that each set of subscriptions cover a corresponding area of event space). If a set of subscriptions has poor locality, the imprecise summary of the set of subscriptions will result in more false positive event traffic than if the set of subscriptions has good locality. Providing for good subscription locality further enhances the throughput of the event distribution network as a whole. Good subscription locality is maintained by routing new subscriptions to the matcher node with the subscription summary that best covers the new subscription.

[0014] Event space partitioning is sometimes desirable but event space partitioning is used in circumstances where subscription locality isn't applicable in the same way. When event space partitioning is desirable, the event space is over-partitioned and a set of event space partitions is assigned to each matcher node in order to provide for more fine-grained load balancing without repartitioning and, ultimately, to provide for enhanced event distribution network throughput, particularly when combined with event routing using imprecise summaries.

Problems solved by technology

If the event router node is overloaded, it reduces the precision of the imprecise summaries.

False positive traffic makes a matcher node work harder.

If a set of subscriptions has poor locality, the imprecise summary of the set of subscriptions will result in more false positive event traffic than if the set of subscriptions has good locality.

Images