On 01/08/2013 09:52 AM, Marek Lindner wrote:
> 
> Hi,
> 
> you don't like the mailing list ? Why do you keep mailing me privately ? this 
> is the last time I respond to mails you send to me privately concerning 
> batman-adv. Use the mailing list!
> 
> On Tuesday, January 08, 2013 19:36:45 you wrote:
>> I am not comparing the ping between two nodes.
>>
>> I am just concerned about the ping latency between nodes.
>>
>> As per my experiment, I am getting 100ms delay between one hop(SPK3
>> –to-SPK2) and 200 ms delay between two hops (SPK3-to-SPK1) and so on.
>>
>> Even 100ms delay seems a lot, some applications may not work properly.
>>
>> Is there any delay (buffer) involved in local packet delivery from  bat0 to
>> local mesh interfaces (wlan0)?
> 
> yes, there is standard wifi buffering. Consider this:
> 
> ping time from node1 to node2: ~100ms
> ping time from node2 to node3: ~100ms
> 
> => ping time from node1 to node3: ~200ms
> 
> The delays simply add up because wifi is a shared medium. The node in the 
> middle can only forward the packet while it does not receive. That is due to 
> how wifi works and has nothing to do with batman-adv.
> 
> If you wish to mitigate the delay you should not use a shared medium. For 
> example using multiple wifi cards on different channels is a solution.


There must be some other problem involved here. We have very good ping
times with batman advanced in working networks as well as test setups.
See this article for example:

http://blog.altermundi.net/article/multiple-hop-mesh-performance-with-multi-band-dual/



Reply via email to