[Battlemesh] tests update

[NicoEchániz]

On 04/23/2013 01:25 PM, Clauz wrote:
> On 04/21/2013 08:41 AM, Axel Neumann wrote:
>> On 04/19/2013 06:13 PM, Clauz wrote:
>>> Hi, battlers.
>>> The mesh is currently set on 54Mbps multicast rate, up to date and
>>> running, after some bugfixes.
>>>
>>> We have some real-time topologies:
>>> https://gates.homelinux.org/ffmap/nodes.html
>>> http://mapa.mesh.altermundi.net/
>>> (select WBMv6 Aalborg on the select box on the top left of the page)
>>>
>>> The test raw data (to be processed) is being uploaded here:
>>> http://downloads.battlemesh.org/WBMv6/test_data/
>>
>> Wow, seems like this year the testbed allowed significantly improved
>> performance (compared to all previous years).
>> For example between up to 4 MBit throughput over eight or more hops in
>> the netperf measurements.
>> Or "only" about 25% percent packet loss in a mobile node scenario with
>> up to 11 hops!!!
>>
>> One request, please, please, if there is any chance to capture some
>> protocol-traffic overhead.
>> This year is also the first time that IPv6 and multiradio was involved
>> and I would be very interested to get an idea of the consequences with
>> 4-times larger address spaces. I think it would be enough to do this
>> captures from a notebook near on of the nodes for each protocol and for
>> only a few minutes.
>>
>>
>> Anyway, Big congratulations and lots of thanks for all the efforts you did!
> 
> Thanks, Axel.
> It's a pity you weren't at this battlemesh!
> 
> I begun documentation, cleanup and processing of the data here:
> https://github.com/cl4u2/-battlemeshv6_tests_postprocessing
> 
> I think that also NicoEchaniz is analysing the data.
> 
> I'll try to work on it on my spare time (if any :))

I quickly parsed the netperf tests output on the same day but I haven't
been in one place more than a day since then. I'll also try to work on
it when I settle for a couple of days in Catalunya before the WCW.

Would it be of any use to have the testbed (or part of it) set up during
the WCW in Berlin to try to run the tests that were left out?

cheers,
NicoEchániz