[srslte-users] Unexpected traffic on LimeSDRmini

[Andrea Valori]

Dear Andre,
Now that I am back in the lab I tested the Lime with the "new" srsRAN (the
experiments I was referring to were done with the older version of srsLTE).
At the moment I am unable to replicate the issue (which may make my issue
obsolete).
The traffic we did see was also at the attach (before starting any "heavy
transmission" with iperf). It might be interesting to discover why Lime and
B200 were performing so differently (especially considering your statement
about no relationship). However, if I will not be able to see the same with
the latest srsRAN version it would be probably a waste of time.

Thanks

On Mon, 30 Aug 2021 at 10:50, Andre Puschmann <andre at srs.io> wrote:

> On 12/8/21 12:20, Andrea Valori wrote:
> > Dear All,
> >
> > We are comparing srsRAN run on B200 and LimeSDRmini.____
> >
> > The Lime has been very satisfactory, especially considering the gap in
> > price compared to the B200. ____
> >
> > However, something strange is happening. Using the Lime, just after the
> > attach, and after stopping full bandwidth stimulation (using iperf) the
> > trace of srsenb keeps showing max bitrate for some time after the end of
> > the stimulation. This does not seem to happen if we are using the B200
> > instead. With the B200 there is minimal traffic generated by the attach
> > procedure, but the full traffic starts only when we start iperf. Also,
> > as soon as iperf ends, the traffic drops to almost nothing immediately.
> >
> > Did someone see this effect? If so, did you understand why?
>
> Hey,
>
> perhaps there is still some traffic buffered when you stop the iperf and
> this buffer will be flushed. Not sure why this happens when you use the
> Lime, perhaps the max rate isn't achieved.
>
> But there is no relationship between the RF front end and the traffic
> generation, etc.
>
> Thanks
> Andre
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.srsran.com/pipermail/srsran-users/attachments/20210830/5f7a93b5/attachment.htm>