In looking for ways to push QoS out throughout a large, distributed,
multi-vendor environment, operational overhead is one of the largest
costs, both in people time and potentially capital expense (tools, training, etc.).
Given the cross platform, standard nature of the OpenFlow protocol and its
ability to push policy to large numbers of devices via a controller,
one would wonder, is QoS one of the killer OpenFlow apps? While not
really available in current the openflow protocol, there is talk of
Flow-Service Classification in future releases. At face value it
looks like this may be a great use case for deploying openflow in
networks. What do folks think about this as a concept? Is it enough
to get interest in OpenFlow as a future enterprise campus option?
Sound off.