I can't think of a reason why anyone would deploy a layer 2 mesh with  
an Ethernet based medium (which wifi inherently is).  Conventional  
wisdom in large scale sp architecture is to do anything of any size or  
complexity in layer 3.  Layer 2 is really bad at scalability and  
really hard to troubleshoot compared to layer 3 as layer 2 "routing"  
is inherently quite "dumb".

If you need l2 functionality or protocol agnostic (although the latter  
is more of an academic feature than a practical benefit), then go l3  
and tunnel.  Most l2 services provided by service providers are, in  
the end, tunneled over a layer 3 infrastructure.  Scalabiity and  
stability are the 2 concerns of a service provider, and both are very  
weak at layer 2 of any size..

-Clint Ricker

On Jun 15, 2008, at 21:00, Matt Hardy <[EMAIL PROTECTED]> wrote:

>
> On Sun, 2008-06-15 at 20:52 -0400, Matt Hardy wrote:
>
>> Yes a layer 2 mesh is protocol dependent, so you're stuck to IP  
>> traffic
>> only.
>
>
>
> Oops... i mean, Layer 3 is protocol dependent :)
>
>
>
>> Also, when using a layer 3 mesh, roaming and convergence time can  
>> also
>> increase (slowing things down) as when things move around, extra  
>> things
>> have to happen... layer 3 stuff... OLSR tables updated, IPs  
>> updated, ARP
>> entries updated, etc
>>
>> For instance, if a laptop migrates from one mesh AP to a different  
>> mesh
>> AP in L3, they will be assigned an IP in a different subnet, while  
>> with
>> a Layer 2 mesh, they can use the same IP.
>>
>> -Matt
>>
>> On Sat, 2008-06-14 at 10:08 -0700, Charles N Wyble wrote:
>>
>>> Rogelio wrote:
>>>> Matt Hardy wrote:
>>>>
>>>>> I guess one question would be is it a Layer 2 or Layer 3 me 
>>>>> sh? That
>>>>> would influence what options you have.
>>>>>
>>>>
>>>> Good question.  Thus far, I've only played with "layer 2" meshes.
>>>> (MobileIP is, I believe, a "layer 3" one, right?)
>>>>
>>>
>>> Yes that is correct.
>>>
>>>> (Layer 2 meshes, I have heard from others, are "better", but I'm  
>>>> not
>>>> exactly sure why this is the case, to be honest.)
>>>>
>>>
>>> Well. It's completely transparent and application/protocol  
>>> independent.
>>>
>>> Charles
>>>
>>>
>>>
>>> --- 
>>> --- 
>>> --- 
>>> --- 
>>> --------------------------------------------------------------------
>>> WISPA Wants You! Join today!
>>> http://signup.wispa.org/
>>> --- 
>>> --- 
>>> --- 
>>> --- 
>>> --------------------------------------------------------------------
>>>
>>> WISPA Wireless List: [email protected]
>>>
>>> Subscribe/Unsubscribe:
>>> http://lists.wispa.org/mailman/listinfo/wireless
>>>
>>> Archives: http://lists.wispa.org/pipermail/wireless/
>>
>>
>> --- 
>> --- 
>> --- 
>> --- 
>> --------------------------------------------------------------------
>> WISPA Wants You! Join today!
>> http://signup.wispa.org/
>> --- 
>> --- 
>> --- 
>> --- 
>> --------------------------------------------------------------------
>>
>> WISPA Wireless List: [email protected]
>>
>> Subscribe/Unsubscribe:
>> http://lists.wispa.org/mailman/listinfo/wireless
>>
>> Archives: http://lists.wispa.org/pipermail/wireless/
>
>
> --- 
> --- 
> --- 
> --- 
> --------------------------------------------------------------------
> WISPA Wants You! Join today!
> http://signup.wispa.org/
> --- 
> --- 
> --- 
> --- 
> --------------------------------------------------------------------
>
> WISPA Wireless List: [email protected]
>
> Subscribe/Unsubscribe:
> http://lists.wispa.org/mailman/listinfo/wireless
>
> Archives: http://lists.wispa.org/pipermail/wireless/


--------------------------------------------------------------------------------
WISPA Wants You! Join today!
http://signup.wispa.org/
--------------------------------------------------------------------------------
 
WISPA Wireless List: [email protected]

Subscribe/Unsubscribe:
http://lists.wispa.org/mailman/listinfo/wireless

Archives: http://lists.wispa.org/pipermail/wireless/

Reply via email to