I think this has come up a number of times in the past few years on nox-dev, pox-dev, and openflow-discuss. For example, see:
https://www.mail-archive.com/pox-dev@lists.noxrepo.org/msg01244.html The basic idea is that you sample the counters at time A and you sample them at time B, subtract the first from the second and then divide by the time between samples. The biggest gotcha here is that it assumes the traffic to and from the switch for the counter request/response, the processing in the switch, and the processing in the controller take equal time for each sample. Depending on your needs for accuracy, you may want to implement a strategy for detecting/mitigating the variance. Alternatively, if your switch supports some other monitoring system, it may be more suitable for this task. For example, IPFIX puts a timestamp in the report packets, so variance in communication time between switch and controller and within the controller don't matter. -- Murphy On Oct 28, 2014, at 12:16 AM, saeedeh karami <saeedehkaram...@gmail.com> wrote: > hi all, > i have some information about flow on links(for example, links between s1 > and s2) by run of flow-state-request function in pox, now i want know, how > can i calculate traffic rate with this data? > can any body help me? > > On Mon, Oct 27, 2014 at 4:47 PM, saeedeh karami <saeedehkaram...@gmail.com> > wrote: > > hi all, > i have some information about flow on links(for example, links between s1 > and s2) by run of flow-state-request function in pox, now i want know, how > can i calculate traffic rate with this data? > can any body help me? > -- > SaeedehKarami > > > > -- > SaeedehKarami
