Methods, apparatus, and systems for routing information flows in networks using spanning trees and network switching element resources

摘要

Methods, apparatus and systems for routing information flows in networks based on spanning trees and network switching element resources. One or more controllers are used to assign information flows to network switching elements (NSEs) through use of spanning trees derived from link path costs. NSEs generate status information relating to resources they employ to facilitate information flows that is sent to the controller(s). The status information is used to derive link costs, which are then used to generate spanning trees that support routing between the NSEs without any path loops. Information flows are assigned to the NSEs such that the routing paths for the flows use the links in the spanning tree. The link costs and spanning trees are dynamically computed during ongoing operations, enabling the network routing and flow assignments to be reconfigured in response to dataplane events and changes to the information flow traffic.

主权项

1. A method for routing information flows in a network including a plurality of network switching elements (NSE) and a plurality of links, each NSE communicatively coupled via at least one link to at least one other NSE and communicatively coupled to a controller, the method comprising:
receiving status information from the NSEs, the status information including data relating to processor resources used by the NSEs to facilitate routing of information flows assigned to the NSEs; determining link path costs for routing paths between the NSEs in the network using data including the data relating to the processor resources in the status information for the NSEs; computing a spanning tree comprising a link path routing tree that enables each NSE to reach each other NSE and has no looped paths, the link paths in the spanning tree determined as a function of the link path costs; and assigning information flows to the NSEs in the network so as to route information flows via the link paths in the spanning tree, wherein the data relating to processor resources in the status information for an NSE includes a capability information and utilization information for at least one processor in the NSE.