Some community wireless / WDS feasibility questions

[Tyler van Houwelingen]

Mixing and Matching vendors is (IMHO) not the right way to go for 
infrastructure.  There is way too much that can cause issues even within the 
same vendor.  I believe the solution is to specificy (or create) a single, 
cheap hardware box that can do "auto-routed-mesh" (which is much more than 
WDS bridging).  The wrt54g was a great starting point with the hacked 
firmware.  The newer WRTs are cheap and much more reliable.  Google is using 
this for their FON box.

The real trick, however, is that the mesh needs to be a routed mesh, not 
bridged WDS or routed with multiple subnets - Otherwise you 1. can only have 
one internet connection feeding it or 2. users have no mobility and will 
constatly be flopping between subnets, forcing DHCP renews if you make the 
network pervasive, 3. Any user can bring down network with a single network 
Dup, ethernet loop or backwards DHCP server plugged in.  The only ones I 
know that have truly accomplished this to date is Tropos but that is quite 
pricey.  Cisco has a great solution, but not for multiple Internet drains as 
it is controller based.

An example of this approach is Meraki, with their potential for success 
coming from  1. Google as an investor always helps, 2. More importantly the 
fact that they are shaping completely the hardware/software and attempting 
to do the above routed mesh.

I would start with the WRT54G and work to make it a true routed mesh. 
(Locustworld may be another good starting point.)

ty

--------
Tyler van Houwelingen
Founder and Chairman
Azulstar, Inc.
1051 Jackson, Grand Haven, MI 49417
Main:  1-877-AZULSTAR
Fax:  616-842-1104
www.azulstar.com

----- Original Message ----- 
From: "Fred Weston" <fweston at gmail.com>
To: <george at awarenessworkshop.org>; "SeattleWireless Talk List" 
<talk at seattlewireless.net>
Sent: Friday, March 09, 2007 10:13 AM
Subject: Re: Some community wireless / WDS feasibility questions


> Bigger than that, but I don't really know how big it could eventually
> become.  It will look more like a line than a cloud though.  I would
> be more or less in the middle of the line and houses around me would
> extend it out toward the rest of the neighborhood.
>
> On 3/8/07, George Lathrop <george at awarenessworkshop.org> wrote:
>> How big is this network? I live in a 3 house compound and we all share a
>> an Internet connection. We do not share a network, altho we could. NO
>> repeaters involved as the distance -100ft. Every client has a different
>> brand of client. Standards are great.
>>
>> Fred Weston wrote:
>> > So, the wireless bug's been getting me again as of late.  I've been
>> > reading up on WDS and looking at cheapish used hardware on eBay trying
>> > to figure out what it would cost to set up a small community network.
>> > Anybody have any input as to the feasibility of WDS for that type of
>> > thing?  I realize WDS means lower overall bandwidth but that isn't
>> > necessarily a huge problem for me.
>> >
>> >>From what I've read, it looks like some manufacturers use WDS that
>> > only works with the exact same model AP throughout the network... are
>> > there any that are more standardized so hardware could be mixed?  Any
>> > thoughts on specific models to look for (cost is the primary factor,
>> > but then again I don't want to spend money on garbage)...?  I guess I
>> > really just want to hear some stories from people with real world
>> > experience with a similar scenario; does it work / not work, what
>> > problems did you run into, etc.
>> >
>> > Thanks
>> > _______________________________________________
>> > Talk mailing list
>> > Talk at seattlewireless.net
>> > http://seattlewireless.net/mailman/listinfo/talk
>> >
>> >
>> >
>> _______________________________________________
>> Talk mailing list
>> Talk at seattlewireless.net
>> http://seattlewireless.net/mailman/listinfo/talk
>>
> _______________________________________________
> Talk mailing list
> Talk at seattlewireless.net
> http://seattlewireless.net/mailman/listinfo/talk
>