Hi Yile,

> On Thu, Apr 12, 2012 at 5:08 PM, Thomas Pedersen <[email protected]> wrote:
>> If B never sent a peering close when shutting down, A will tear down
>> the peer link after ~15 failed transmissions.

On Thu, Apr 19, 2012 at 8:33 AM, Yile Ku <[email protected]> wrote:
> So if we turn off a node it will take 15 seconds to
> find another path..?

15 failed transmissions only equals 15 seconds if you are sending 1
data frame per second.
In our tests with audio streams the path switch takes ~250ms, and that
is with the default mesh configuration parameters.  You could tune
your network to do better than that.

Cheers,

Javier



> On Thu, Apr 12, 2012 at 5:08 PM, Thomas Pedersen <[email protected]> wrote:
>>
>> Hi Yile,
>>
>> On Thu, Apr 12, 2012 at 3:26 PM, Yile Ku <[email protected]> wrote:
>> > I have 4 nodes A,B,C, & D.  I have A set to not directly connect to D.
>> >  Attached to D I have a web based camera.  I have a laptop hooked to
>> > node A that I can log into the camera using a browser.    I can
>> > connect to the camera okay going through node B.  I shut off node B
>> > and the video session dies.  I assume that 802.11s would switch the
>> > route of traffic to go through node C, but it doesn't appear to do
>> > that.  I tried settting the root node flag on both A and D but that
>> > did not help.
>> >
>> > I tried pinging directly from A to the camera attached to D.  Here is
>> > a list of ping times:
>> > 64 bytes from 192.168.0.224: seq=63 ttl=64 time=1.950 ms
>> > 64 bytes from 192.168.0.224: seq=64 ttl=64 time=1.992 ms
>> > 64 bytes from 192.168.0.224: seq=65 ttl=64 time=2.877 ms
>> > 64 bytes from 192.168.0.224: seq=66 ttl=64 time=1.963 ms
>> > 64 bytes from 192.168.0.224: seq=67 ttl=64 time=526.659 ms
>> > 64 bytes from 192.168.0.224: seq=68 ttl=64 time=6.892 ms
>> > 64 bytes from 192.168.0.224: seq=69 ttl=64 time=1.979 ms
>> > 64 bytes from 192.168.0.224: seq=70 ttl=64 time=2.867 ms
>> > 64 bytes from 192.168.0.224: seq=71 ttl=64 time=1.968 ms
>> > 64 bytes from 192.168.0.224: seq=72 ttl=64 time=1.961 ms
>> > 64 bytes from 192.168.0.224: seq=77 ttl=64 time=506.990 ms
>> > 64 bytes from 192.168.0.224: seq=78 ttl=64 time=3.043 ms
>> > 64 bytes from 192.168.0.224: seq=79 ttl=64 time=1.954 ms
>> > 64 bytes from 192.168.0.224: seq=80 ttl=64 time=3.768 ms
>> > 64 bytes from 192.168.0.224: seq=81 ttl=64 time=3.913 ms
>> > 64 bytes from 192.168.0.224: seq=82 ttl=64 time=4.725 ms
>> > 64 bytes from 192.168.0.224: seq=83 ttl=64 time=57.908 ms
>> > 64 bytes from 192.168.0.224: seq=84 ttl=64 time=6.435 ms
>> > 64 bytes from 192.168.0.224: seq=85 ttl=64 time=1.968 ms
>> > 64 bytes from 192.168.0.224: seq=86 ttl=64 time=1.974 ms
>> > 64 bytes from 192.168.0.224: seq=87 ttl=64 time=1.956 ms
>> > 64 bytes from 192.168.0.224: seq=88 ttl=64 time=1.960 ms
>> > 64 bytes from 192.168.0.224: seq=89 ttl=64 time=1.994 ms
>> > 64 bytes from 192.168.0.224: seq=90 ttl=64 time=2.010 ms
>> > 64 bytes from 192.168.0.224: seq=91 ttl=64 time=2.848 ms
>> > 64 bytes from 192.168.0.224: seq=92 ttl=64 time=22.417 ms
>> > 64 bytes from 192.168.0.224: seq=93 ttl=64 time=3.196 ms
>> > 64 bytes from 192.168.0.224: seq=94 ttl=64 time=1.975 ms
>> > 64 bytes from 192.168.0.224: seq=95 ttl=64 time=12.603 ms
>> > 64 bytes from 192.168.0.224: seq=96 ttl=64 time=9.577 ms
>> > 64 bytes from 192.168.0.224: seq=97 ttl=64 time=2.065 ms
>> > 64 bytes from 192.168.0.224: seq=98 ttl=64 time=9.883 ms
>> > 64 bytes from 192.168.0.224: seq=99 ttl=64 time=8.772 ms
>> > 64 bytes from 192.168.0.224: seq=100 ttl=64 time=8.407 ms
>> > 64 bytes from 192.168.0.224: seq=101 ttl=64 time=7.967 ms
>> > 64 bytes from 192.168.0.224: seq=102 ttl=64 time=9.332 ms
>> > 64 bytes from 192.168.0.224: seq=103 ttl=64 time=6.949 ms
>> > ^C
>> > --- 192.168.0.224 ping statistics ---
>> > 104 packets transmitted, 72 packets received, 30% packet loss
>> > round-trip min/avg/max = 1.917/32.764/526.659 ms
>> > root@OpenWrt:/# ping 192.168.0.224
>> > PING 192.168.0.224 (192.168.0.224): 56 data bytes
>> > 64 bytes from 192.168.0.224: seq=0 ttl=64 time=2.238 ms
>> > 64 bytes from 192.168.0.224: seq=1 ttl=64 time=1.965 ms
>> > 64 bytes from 192.168.0.224: seq=2 ttl=64 time=7.304 ms
>> > 64 bytes from 192.168.0.224: seq=3 ttl=64 time=2.824 ms
>> > 64 bytes from 192.168.0.224: seq=8 ttl=64 time=425.330 ms
>> > 64 bytes from 192.168.0.224: seq=9 ttl=64 time=3.783 ms
>> > 64 bytes from 192.168.0.224: seq=10 ttl=64 time=2.353 ms
>> > 64 bytes from 192.168.0.224: seq=11 ttl=64 time=2.337 ms
>> >
>> >
>> > The ping times vary greatly.  why?
>>
>> This is likely related to the path refresh timer. While an mpath is
>> refreshing, it is marked as deactivated. We try to pre-empt this and
>> it seems to work most of the time, but perhaps starting the path
>> refresh 1 second before deactivating is not enough time? See
>> mesh_nexthop_lookup() and hwmp_rann_frame_process().
>>
>> > Also how can I get it to switch fast from B to C when routing from A to
>> > D?
>>
>> Does A never switch from B to C, or does it just take a while?
>> If B never sent a peering close when shutting down, A will tear down
>> the peer link after ~15 failed transmissions. Depending on the type of
>> traffic, this might take a while. See ieee80211s_update_metric() in
>> net/mac80211/status.c
>>
>> Patches welcome.
>>
>> Thomas
>> _______________________________________________
>> Devel mailing list
>> [email protected]
>> http://lists.open80211s.org/cgi-bin/mailman/listinfo/devel
>
>
>
> _______________________________________________
> Devel mailing list
> [email protected]
> http://lists.open80211s.org/cgi-bin/mailman/listinfo/devel
>



-- 
Javier Cardona
cozybit Inc.
http://www.cozybit.com
_______________________________________________
Devel mailing list
[email protected]
http://lists.open80211s.org/cgi-bin/mailman/listinfo/devel

Reply via email to