Hi Linus, Simon,

Testing with static routes yields the same result.  After thorough
testing, i found strange throughput pattern on one of my radio.  It's
my hardware problem after all.  Phew...
Thanks a lot for your input. I appreciate it.

Rgds,
Jack.

On Thu, Oct 17, 2013 at 12:28 PM, Linus Lüssing <[email protected]> wrote:
> Hi Jack,
>
> One more test could be interesting: What throughput do you get
> with static routes over the wifi interfaces directly, without
> batman-adv?
>
> Just to check whether it's batman-adv's fault or something else
> (e.g. cheap, unshielded wifi chips, a bug in the test setup /
> configuration).
>
> Cheers, Linus
>
>
> On Thu, Oct 17, 2013 at 11:21:57AM +0800, Zaki wrote:
>> Hi Linus,
>>
>> From A to B(radio 1) it is 5.24GHz. Throughput is 40Mbps.
>>
>> From B(radio 2) to C it is 2.412GHz.
>> Throughput is 38Mbps.
>>
>> From A to C the throughput is just 20Mbps.
>>
>> Thanks.
>> Jack.
>>
>> On Mon, Oct 14, 2013 at 11:37 PM, Linus Lüssing <[email protected]> 
>> wrote:
>> > Hi Jack,
>> >
>> > what is the throughput you get from B to C? What was the
>> > throughput you got on the 2.4GHz link alone? Did you change to
>> > 2.4GHz on A-B or B-C?
>> >
>> > Cheers, Linus
>> >
>> >
>> > On Mon, Oct 14, 2013 at 09:26:41AM +0800, Zaki wrote:
>> >> Hi Simon,
>> >>
>> >> Thanks for your reply.  I am using omni antenna right now and what i
>> >> got when i did an iPerf from A to C is only 20Mbps.  It looks like the
>> >> throughput is cut to half (just like a single radio mesh).  Tried 2.4G
>> >> frequency in one of the pair, had the same result too.  Anything in
>> >> terms of setting of batman-adv i can look for?.
>> >>
>> >> Rgds,
>> >> Jack
>> >>
>> >> On Sun, Oct 13, 2013 at 10:36 PM, Simon Wunderlich
>> >> <[email protected]> wrote:
>> >> > Hi Zaki/Jack,
>> >> >
>> >> > On Fri, Oct 11, 2013 at 11:11:04AM +0800, Zaki wrote:
>> >> >> Hi,
>> >> >>
>> >> >> I am a newbie in batman-adv.  I got a question regarding Hops and
>> >> >> Throughput, especially in regards with a collocated dual-radio device.
>> >> >>  Hopefully somebody could offer some help here for me to fully
>> >> >> understand batman-adv operation.  I have 3 devices, namely A, B and C.
>> >> >>
>> >> >> A is only having single radio (operating at Freq 5.20G in AdHoc mode).
>> >> >>  Put this under bat0.  Bridge bat0 with eth0.
>> >> >> B is having dual-radio (radio1 operates at 5.20G in AdHoc mode, radio2
>> >> >> is at 5.24G in AdHoc mode). Put these under bat0. Bridge bat0 with
>> >> >> eth0
>> >> >> C is like A (single radio, operating at 5.24G in AdHoc mode). Put this
>> >> >> under bat0.  Bridge bat0 with eth0.
>> >> >>
>> >> >> So, the wireless link will be like this :
>> >> >>
>> >> >> [A]::::::::::::::: [B(radio1)---B(radio2)]::::::::::::::::[C]
>> >> >>
>> >> >> 1) If i were to ping from A to C, how many hops is it considered?.  Is
>> >> >> it 2 hops or 3 hops?.
>> >> >
>> >> > that would be two hops. One hop is considered traversing one link, and
>> >> > you have two links here.
>> >> >
>> >> >> 2) If my Iperf TCP throughput from A to B is 40Mbps, what is the
>> >> >> expected throughput of Iperf from A to C ?.
>> >> >
>> >> > That depends. If the links are not interferring, you should get approx.
>> >> > the maximum of both link throughputs. So if one link is 30 mbit/s, the 
>> >> > second
>> >> > is 25 mbit/s, that whole path would yield 25 mbit/s - there might be tcp
>> >> > effects further decreasing throughput (latency etc).
>> >> >
>> >> > However you use two frequencies adjacent to each other. If you use omni
>> >> > antennas, expect a lot of interferences (even if the channels are not
>> >> > overlapping). If you use directional antennas you should be fine.
>> >> >
>> >> > Cheers,
>> >> >         Simon
>> >> >
>> >> >
>> >> >>
>> >> >> Thanks a lot.
>> >> >>
>> >> >> Jack.

Reply via email to