[srsran-users] eMBMS throughput limit?

[Justin Tallon]

Hey Joe & Ron,

Apologies for the delayed response, I have just now seen your emails.
With the default configuration, you should be able to get 11MS/s over the
MBMS. This is using  PRB=50 and embms.mcs=20, as well as the maximum
assignment of subframes (6/10).
If you increase the MCS=24(maximum allowable for eMBMS, due to the greater
number of REs dedicated to reference signals) and set PRB=50, you should
get 16MS/s.
I have tested this using ZMQ on Ubuntu 21.04. I have noticed an issue with
100prbs, not the crashing as you described but some PHY layer KOs, I will
investigate this and have a fix for the next release.

Ron, can you share the iperf command you are using? It's possible you are
flooding the buffers in the RLC which can throw off the eMBMS scheduler,
causing a drop in performance.

Joe, I will investigate the issue you have created and get back to you.

Thanks!
Justin

Justin Tallon

Engineering Manager

Software Radio Systems
justin at srs.io
www.srs.io
Barcelona, Spain
[image: twitter] <https://twitter.com/srssystems>
[image: linkedin] <https://www.linkedin.com/in/justin-tallon-71807136/>


On Thu, Jul 28, 2022 at 4:16 PM J Giovatto <jgiovatto at adjacentlink.com>
wrote:

> Hi Ron,
>
> This question probably belongs on the srsRAN-emane page but with that
> being said maybe the kind folks from srs can chime in and
>
> report back any thruput testing they may have done with multicast.
>
> One thing you might want to try is the zmg rf and see if you get different
> results vs emane to help determine if this is an srs or emane porblem.
>
> Also note there is an issue related to multicast here:
> https://github.com/srsran/srsRAN/issues/936
>
> Lastly your versions are looking out of date, we are on 21.10 and only
> supporting Ubuntu 20.04 so any testing will not be a true apples to apples.
>
> Joe
>
>
> On 7/27/22 19:25, Sackman (US), Ronald W wrote:
>
> Hello,
>
>
>
> I am running SRSRAN version 19.9.0 in EMANE, running on Ubuntu 18.04.  I
> followed the
>
> instructions in the user manual section 5.12 “eMBMS End-to-end” and (with
> a few modifications)
>
> was able to get it working.  Testing with iperf or MGEN, I am finding that
> the multicast throughput
>
> is limited to 6Mbps, with an additional 6Mbps of throughput allowed for
> unicast.  If I
>
> attempt to increase throughput by increasing the enb.conf/p_nrb parameter
> above
>
> 25 (e.g. to 50 or 100), the receiving srsue interface crashes as soon as
> it receives the
>
> first packet.
>
>
>
> Using the following calculator with SRSRAN MCS=26.5 (256QAM), n_prb=25,
> and 1X1 SiSo, I am
>
> expecting a total of 24.48Mbps downlink bandwidth, out of which I assume
> multicast would
>
> consume some carve out without reducing the total cell bandwidth.
> However, the cell maxes
>
> out at 6Mbps multicast + 6Mbps unicast.
>
> https://www.cellmapper.net/4G-speed
>
>
>
> Is this the expected throughput performance?  Am I missing a step to
> achieve the calculated
>
> throughput for the cell?  I have attached the ENB and UE config files.
>
>
>
> Thank you,
>
> Ron
>
>
>
> =======================================
> Ron Sackman
>
> Technical Fellow
> Principal Network Security Architect
> Boeing Autonomous Systems / Critical Systems Security
>
> Mountain View CA
> Office: 650-316-3749 Cell: 408-930-7358
> ronald.w.sackman at boeing.com
>
> *Psalm 19:1*
>
>
>
> _______________________________________________
> srsran-users mailing listsrsran-users at lists.srsran.comhttps://lists.srsran.com/mailman/listinfo/srsran-users
>
>
> _______________________________________________
> srsran-users mailing list
> srsran-users at lists.srsran.com
> https://lists.srsran.com/mailman/listinfo/srsran-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.srsran.com/pipermail/srsran-users/attachments/20220906/1b80edbf/attachment-0001.htm>