Hello, The new flowvisor does not yet have this feature, but it's on the short list of things for me to do ASAP.
You can follow the progress at: https://openflow.stanford.edu/bugs/browse/FLOWVISOR-81 - Rob . On Sun, May 15, 2011 at 6:03 PM, <[email protected]> wrote: > Hello Yiannis, > > I was really confusing the subjects. Now I understood the parameter > priority. > > Just another question. I think that in the old version of FlowVisor I could > configure a Bandwidth parameter. I was thinking that priority parameter has > replaced it. Is there some parameter to add a simple QoS mechanism like > bandwidth in FlowVisor? > > Thank you. > > My best regards. > > Luciano Ignaczak > > Citando Yiannis Yiakoumis <[email protected]>: > >> Hi Luciano, >> >> I think you are confusing priority of a flow-entry with QoS-related >> priority. >> Priority will define which flow-entry the incoming packets will match. In >> your case, traffic that goes to tp_dst=5060 matches both slices >> definition, >> but it belongs (and therefore controlled) by the one with highest >> priority. >> It doesn't have any impact on the rates these flows will achieve. >> >> For this you might want to look at min-rate queues : >> http://www.openflow.org/wk/index.php/Slicing >> >> Regards, >> Yiannis >> >> On Sat, May 14, 2011 at 11:26 PM, <[email protected]> wrote: >> >>> Hello, >>> >>> I´d like a little help to know if I understood correctly the use of >>> priority parameter on a FlowSpace. >>> >>> I created a topology and divided it in two slices: voice and data. After >>> this I configured two FlowSpaces on each switch following the examples >>> below: >>> fvctl addFlowSpace <MAC> 10000 "tp_dst=5060" "Slice:voice=4" >>> fvctl addFlowSpace <MAC> 1 any "Slice:data=4" >>> >>> When I was reading "man fvctl" I understood that the priority parameter >>> determines the priority of the flow. However, during my tests with Iperf >>> I >>> have got better results with the data´s slice. >>> >>> Am I using correctly the parameter priority? >>> >>> I used mininet + NOX + flowvisor in my tests. >>> >>> Best regards. >>> >>> Luciano Ignaczak >>> >>> ---------------------------------------------------------------- >>> This message was sent using IMP, the Internet Messaging Program. >>> >>> >>> _______________________________________________ >>> openflow-discuss mailing list >>> [email protected] >>> https://mailman.stanford.edu/mailman/listinfo/openflow-discuss >>> >> > > > > ---------------------------------------------------------------- > This message was sent using IMP, the Internet Messaging Program. > > > _______________________________________________ > openflow-discuss mailing list > [email protected] > https://mailman.stanford.edu/mailman/listinfo/openflow-discuss > _______________________________________________ openflow-discuss mailing list [email protected] https://mailman.stanford.edu/mailman/listinfo/openflow-discuss
