b.a.t.m.a.n.lists.open-mesh.org archive mirror
 help / color / mirror / Atom feed
* Re: [B.A.T.M.A.N.] Advice for dense meshes
@ 2013-01-23  9:03 Pau Koning
  0 siblings, 0 replies; 5+ messages in thread
From: Pau Koning @ 2013-01-23  9:03 UTC (permalink / raw)
  To: steve; +Cc: The list for a Better Approach To Mobile Ad-hoc Networking

> I am seeking some general advice on how to optimise batman-adv in the
> context of a dense mesh.

Update to newest compat wireless
Change multicast rate to something higher like 18M/54M
Change radio distance setting to something < 300m
Increase beacon_int of all VIf to something like 5000 (5 sec) or higher
Enable sst in Adhoc (modifying source for 2.4GHz required
http://article.gmane.org/gmane.linux.kernel.wireless.general/100844 )
Increase batman-adv ogm interval to 5000 or 10000
Increase batman-adv hop penalty to to 50 or 70

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [B.A.T.M.A.N.] Advice for dense meshes
  2013-01-22 15:37   ` L. Aaron Kaplan
@ 2013-01-22 17:28     ` Steve Song
  0 siblings, 0 replies; 5+ messages in thread
From: Steve Song @ 2013-01-22 17:28 UTC (permalink / raw)
  To: b.a.t.m.a.n

Hi Jan, Aaron,

On 22 January 2013 11:37, L. Aaron Kaplan <aaron@lo-res.org> wrote:
>
> On Jan 22, 2013, at 1:54 PM, Steve Song <steve@villagetelco.org> wrote:
>
>> Hi all,
>>
>> A colleague of mine is building a batman-adv mesh network in an
>> apartment building with essentially one node per apartment.   Not
>> surprisingly, results in a very dense mesh with each node having a
>> large number of neighbours.  Here is a typical batctl o output
>> http://pastebin.com/aAR43hj7  This results in some fairly slow
>> connections.
>>
>> I am seeking some general advice on how to optimise batman-adv in the
>> context of a dense mesh.  Options that we have considered include
>> turning the radio transmit power down on all of the devices and/or
>> alternating channels on different floors (e.g.  1,6,11,1,6,11,etc).
>> However, it is not clear to us what is the best strategy in this
>> context.  Grateful for any tips or suggestions you may have.
>>
>> Thanks in advance.... Steve Song
>>
>
> Hi Steve,
>
> The following is my experience.
>
> Optimise layer 1 by:
> *) proper channel assignments
> *) reduce txpower
> *) dont's send when you don't have to send
> *) use spacial separation of frequencies and directionality if possible (ok, this is more of a general advice, won't matter so much in apartments)
> *) most importantly: use 802.11n!

Many thanks both for the practical advice.

For better or worse we are constrained by the fact that this is a Mesh
Potato network which obliges us to put a node in each apartment in
order to provide phone services via the FXS port in the Mesh Potatoes.
 This is not really the kind of network the Mesh Potato was originally
designed for but was chosen for the apartment building over copper
wiring because of the wiring constraints specific to the building.  I
hope we'll be able to optimise this network for adequate performance.

Regards... Steve

-- 
Steve Song
+1 902 529 0046
+27 83 482 2088 (SMS only)
http://villagetelco.org

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [B.A.T.M.A.N.] Advice for dense meshes
  2013-01-22 12:54 ` Steve Song
  2013-01-22 13:27   ` Jan Lühr
@ 2013-01-22 15:37   ` L. Aaron Kaplan
  2013-01-22 17:28     ` Steve Song
  1 sibling, 1 reply; 5+ messages in thread
From: L. Aaron Kaplan @ 2013-01-22 15:37 UTC (permalink / raw)
  To: steve, The list for a Better Approach To Mobile Ad-hoc Networking


On Jan 22, 2013, at 1:54 PM, Steve Song <steve@villagetelco.org> wrote:

> Hi all,
> 
> A colleague of mine is building a batman-adv mesh network in an
> apartment building with essentially one node per apartment.   Not
> surprisingly, results in a very dense mesh with each node having a
> large number of neighbours.  Here is a typical batctl o output
> http://pastebin.com/aAR43hj7  This results in some fairly slow
> connections.
> 
> I am seeking some general advice on how to optimise batman-adv in the
> context of a dense mesh.  Options that we have considered include
> turning the radio transmit power down on all of the devices and/or
> alternating channels on different floors (e.g.  1,6,11,1,6,11,etc).
> However, it is not clear to us what is the best strategy in this
> context.  Grateful for any tips or suggestions you may have.
> 
> Thanks in advance.... Steve Song
> 

Hi Steve,

The following is my experience.

Optimise layer 1 by:
*) proper channel assignments
*) reduce txpower
*) dont's send when you don't have to send
*) use spacial separation of frequencies and directionality if possible (ok, this is more of a general advice, won't matter so much in apartments)
*) most importantly: use 802.11n!



^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [B.A.T.M.A.N.] Advice for dense meshes
  2013-01-22 12:54 ` Steve Song
@ 2013-01-22 13:27   ` Jan Lühr
  2013-01-22 15:37   ` L. Aaron Kaplan
  1 sibling, 0 replies; 5+ messages in thread
From: Jan Lühr @ 2013-01-22 13:27 UTC (permalink / raw)
  To: steve, The list for a Better Approach To Mobile Ad-hoc Networking

Hello,

Am 22.01.2013 um 13:54 schrieb Steve Song:

> A colleague of mine is building a batman-adv mesh network in an
> apartment building with essentially one node per apartment.   Not
> surprisingly, results in a very dense mesh with each node having a
> large number of neighbours.  Here is a typical batctl o output
> http://pastebin.com/aAR43hj7  This results in some fairly slow
> connections.
> 
> I am seeking some general advice on how to optimise batman-adv in the
> context of a dense mesh.  Options that we have considered include
> turning the radio transmit power down on all of the devices and/or
> alternating channels on different floors (e.g.  1,6,11,1,6,11,etc).
> However, it is not clear to us what is the best strategy in this
> context.  Grateful for any tips or suggestions you may have.
> 

using different channels is always a good idea when it comes to crowded places.
Another option in to use 5 Ghz for mesh and 2.4 Ghz for clients.

Another options is to use less nodes - one node per appartment might be too much.

Turning down the transmit power can help, but might have the opposite effect.
(Interference- vs. transmission range)

Another - imho the most important rule - is to limit hops
Each hop limits the total throughput: if using same wireless channels one hop cannot forward data while receiving. It has to wait.
 
To summarize:
-> Minimize the (expected) hop count.
-> Use as many channels as possible (incl. 5 Ghz)
-> Use as few nodes per channel as needed to provide coverage

Keep smiling
yanosz



^ permalink raw reply	[flat|nested] 5+ messages in thread

* [B.A.T.M.A.N.] Advice for dense meshes
       [not found] <CAD_CWO3cQMtPns+xzc2hs4sUCmZM1ux4-335rhicUn05Upvu-g@mail.gmail.com>
@ 2013-01-22 12:54 ` Steve Song
  2013-01-22 13:27   ` Jan Lühr
  2013-01-22 15:37   ` L. Aaron Kaplan
  0 siblings, 2 replies; 5+ messages in thread
From: Steve Song @ 2013-01-22 12:54 UTC (permalink / raw)
  To: b.a.t.m.a.n

Hi all,

A colleague of mine is building a batman-adv mesh network in an
apartment building with essentially one node per apartment.   Not
surprisingly, results in a very dense mesh with each node having a
large number of neighbours.  Here is a typical batctl o output
http://pastebin.com/aAR43hj7  This results in some fairly slow
connections.

I am seeking some general advice on how to optimise batman-adv in the
context of a dense mesh.  Options that we have considered include
turning the radio transmit power down on all of the devices and/or
alternating channels on different floors (e.g.  1,6,11,1,6,11,etc).
However, it is not clear to us what is the best strategy in this
context.  Grateful for any tips or suggestions you may have.

Thanks in advance.... Steve Song


--
Steve Song
+1 902 529 0046
+27 83 482 2088 (SMS only)
http://manypossibilities.net
http://villagetelco.org

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2013-01-23  9:03 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2013-01-23  9:03 [B.A.T.M.A.N.] Advice for dense meshes Pau Koning
     [not found] <CAD_CWO3cQMtPns+xzc2hs4sUCmZM1ux4-335rhicUn05Upvu-g@mail.gmail.com>
2013-01-22 12:54 ` Steve Song
2013-01-22 13:27   ` Jan Lühr
2013-01-22 15:37   ` L. Aaron Kaplan
2013-01-22 17:28     ` Steve Song

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).