Network clustering

摘要

A network clustering facility is described. The network clustering facility receives a message having a location identification that identifies a destination for the message and comprises at least two components, selects a network node identified by the components of the location identification, determines whether a link exists to the selected network node, and identifies a best link to the selected network node. The network clustering facility comprises a clustering subsystem, an advertising subsystem, and a routing layer.

主权项

1. A method performed by a computing system for message routing in a data communications network, the communications network having a plurality of clusters each having one representative network node and at least one other network node, wherein the representative network node of the cluster is interconnected using links with all other network nodes of its cluster, comprising:
receiving a message with a location identification for the message, the location identification comprising at least a first component that identifies a destination network node for the message and a second component that identifies a representative network node that represents a cluster, and wherein the destination network node is a member of the cluster represented by the representative network node; selecting the representative network node identified by the second component of the location identification; and identifying a best link to the selected representative network node, wherein the best link is identified from routing information in a routing table, wherein the routing table comprises components identifying a subset of a set of network nodes associated with the data communications network; and communicating the message to the representative network node via the identified best link.