Network traffic management

摘要

A virtual entity is used to represent a bandwidth bottleneck point in a network. The virtual bottleneck point may correspond to a collection of components (e.g. routers) or a portion of a real component (e.g. a router could be represented by multiple virtual bottleneck points.) This model allows a user to decide which points in their network require management, independent of the underlying network infrastructure. Bandwidth usage across virtual entities can be tracked and compared to the bandwidth limit available to an application at each bottleneck point. When the bandwidth available at the bottleneck point is fully utilized, additional calls can be blocked or rerouted. Policies can be applied to permit certain calls to proceed despite the fact that bandwidth is fully utilized, or to block certain calls when bandwidth usage is approaching the maximum level. Bandwidth management can be distributed or centralized with information shared throughout a distributed network.