There are a number of bugs in the current 802.11s implementation which
cause poor to disastrous routing (and most of our routing testing has been
with four node networks). We’ve started submitting patches for the bugs we
have fixed but it will take some time to get them all submitted.



Multicast is flooded when the incoming packet doesn’t match a cached
packet. The cache is rather small though. We had evidence that it was too
small and we weren’t using 15 node networks back then. It wouldn’t surprise
me if there were multicast storms with networks that large..




n  Jesse



*From:* Devel [mailto:[email protected]] *On Behalf Of *
[email protected] via Devel
*Sent:* Friday, June 12, 2015 11:47 PM
*To:* [email protected]
*Subject:* Re: Mesh broadcast storm



I also found the mesh works excellent up to 15 nodes. Using hwmp_mode 3
increases the number of nodes to 35 but then the mesh is suffering
performance.



In my case i need to mesh up about 100 nodes in close proximity. Each node
is also an access point and i have only a single radio (carambola2 with
ar9331)



So i am grouping now always about 15 nodes in mesh1 mesh2 mesh3 and had to
use also a seperate channel for each mesh.



Otherwise i get problems with circular routes as the mesh runs crazy in
circular traffic on the backend.



This was my best result so far. Forcing a hierarchical topology improves
the performance at best.



I am interested to learn about other solutions to the topic as this
segmentation is complicated to set up.



------ Original message------

*From: *Francois Gervais via Devel

*Date: *Fri, 12 Jun 2015 18:46

*To: *[email protected];

*Subject:*Mesh broadcast storm



Hi,We have been running a mesh network with 15 nodes for a while now
and it seems stable. We tried upgrading to a 28 nodes mesh and it work
for 20 minutes and then crashed. The mesh still look fine but there
seem to be a storm of ARP who-has broadcast looping around. Is the
concept of broadcast in a mesh network explained in the IEEE
802.11-2012 spec? I'd like to learn more about it to help us diagnose
the problem.Thanks_______________________________________________Devel
mailing [email protected]
<%[email protected]>p://lists.open80211s.org/cgi-bin/mailman/listinfo/devel
_______________________________________________
Devel mailing list
[email protected]
http://lists.open80211s.org/cgi-bin/mailman/listinfo/devel

Reply via email to