OTTAWA, ON. -- Software Defined Networking pioneer, Corsa Technology, the leader in performance SDN switching, today announced new SDN metering and QoS (Quality of Service) capability for its line of performance SDN hardware. Corsa’s SDN implementation of this classic traffic engineering function allows network architects to better manage bandwidth across their network with dynamic, policy-aware metering and QoS. Framed around OpenFlow 1.3 QoS, Corsa performance SDN hardware lets network orchestration dynamically use meters and multiple classes of service to deliver SDN QoS which adjusts and adapts allocation of bandwidth at ultra-granular flow-level.
Metering and queuing allows networks to create bandwidth profiles by putting limits and guarantees on traffic with particular classes. With SDN, those limits are no longer fixed as part of a static topology and rigid hardware platform. Policy-aware provisioning can be dynamically pushed down to the flexible Corsa SDN hardware to make on-going adjustments to meters and queue assignments. The network can then make immediate, informed queuing and discard decisions under congestion. Real-time performance monitoring automatically returns meter statistics and is checked against policy such as SLAs. For network operators including service providers and ISPs, SDN metering and queuing allows new self-serve features to be offered such as “bandwidth reservation” where users can dynamically schedule and reserve bandwidth via separate class of service and meters.
This on-demand bandwidth reservation service is especially interesting for organizations running Big Data workloads. Using Corsa SDN metering and QoS, the network has the ability to respond to a request, set up the appropriate circuits and then recognize Big Data flows. Depending on real-time network conditions, such as traffic levels and congestion, as well as dynamically changing network policy, SDN metering and QoS dynamically routes flows, rate-limits flows and reacts in real-time to Big Data events, such as node addition, node deletion and replication. For the Big Data user, this ensures reliable large data transfers occur predictably in hours instead of days. For the service provider, long cumbersome provisioning cycles are avoided as the service is offered on-demand. Network orchestration can dynamically adjust or adapt policy based on SLAs, types of traffic, types of users, or types of applications and allocation of bandwidth at flow-level can be set up and removed to make sure data transfers occur quickly and without disruption to other users.