[Battlemesh] any (intermediary) results

[Pau]

The script collected 4 completed tests. One test needs one hour (50*5 pings
for each node in the network). I got the latency of each node for each
protocol stored in different files.
So I have all data in my hard disk. Now Axel and I are trying to do it nice,
making a document with graphs and conclusions.
We will put it in WBMv4 web page.

p4u


2011/3/22 Axel Neumann <neumann at cgws.de>

> On Sonntag 20 März 2011, Benjamin Henrion wrote:
> > On Sun, Mar 20, 2011 at 7:02 PM, L. Aaron Kaplan <aaron at lo-res.org>
> wrote:
> > > Hi!
> > >
> > > Back home again (unfortunately). How is the situation at the WBMv4? Did
> > > you manage to run any tests?
> >
> > Axel made some tests early this morning, with big pings, but we had to
> > remove some nodes due to the raising moisture of the morning. The
> > power supply we installed at the italian was also removed to close the
> > door around 6am.
> >
> > No other TCP nor UDP transfers were made AFAIK.
> >
> > Maybe Axel can say more...
>
> We did further tests.
> Where is the recommended place to upload the approximately 50MB of data ?
>
> More detailed:
>
> In the afternoon, Boris and Henning redeploy the nodes that were collected
> the
> previous night because of potential rain. They also put back the 3-channel-
> switched fonera boards back to their old position. In the end we had 34
> active
> nodes in the mesh.
>
> Another problem were the remaining nodes that still had the fixed 54MBit
> rate
> option set.
> IMHO this option only affects unicast transmissions and causes them to
> essentially behaved like a black hole!
> Because of this option, all our previous tests (pings and others unicast
> measurements) were transmitted at 54MBit resulting in a much higher packet
> loss than detected by our routing protocols (which are sending  LinkQuality
> probes based on broadcast/multicast and transmitted with 1 or 5MBit at
> most.
>
> After reenabling the auto rate-selection, the measurements became much more
> plausible. Also we were finally able to ping ALL 34 nodes in the mesh (some
> via
> up to 5 hops and a round-trip packet loss of 80 %).
>
> After that, Henning and me used a script to collect CPU, memory, protocol
> overhead data and ping success rates to various destination nodes (one
> destination node per measurement). We made a bunch of measurements,
> including
> mobile scenarios. Some of the data could need some further post processing
> to
> be better understandable but we already have some nice graphs and numbers.
>
> Pau also wrote an interesting script measuring and visualizing latency and
> ping success rates to ALL nodes and using ALL protcols (static-rotuing,
> batman, babel, olsr, batman-adv, bmx6).
>
> Monday morning at 8 o clock we went to bed, leavin Paus' script running
> until
> the first people woke up and started to clean up everything.
>
>
> /axel
>
> >
> > --
> > Benjamin Henrion <bhenrion at ffii.org>
> > FFII Brussels - +32-484-566109 - +32-2-4148403
> > "In July 2005, after several failed attempts to legalise software
> > patents in Europe, the patent establishment changed its strategy.
> > Instead of explicitly seeking to sanction the patentability of
> > software, they are now seeking to create a central European patent
> > court, which would establish and enforce patentability rules in their
> > favor, without any possibility of correction by competing courts or
> > democratically elected legislators."
> > _______________________________________________
> > Battlemesh mailing list
> > Battlemesh at ml.ninux.org
> > http://ml.ninux.org/mailman/listinfo/battlemesh
>
> _______________________________________________
> Battlemesh mailing list
> Battlemesh at ml.ninux.org
> http://ml.ninux.org/mailman/listinfo/battlemesh
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://ml.ninux.org/pipermail/battlemesh/attachments/20110322/e8590813/attachment-0001.html>