[Sugar-devel] meshing XOs with netbooks (WAS: Re: [Bugs] #79 NORM: SOAS should mesh network when possible)
tomeu at sugarlabs.org
Mon Jun 15 03:39:14 EDT 2009
On Mon, Jun 15, 2009 at 02:32, Gary C Martin<gary at garycmartin.com> wrote:
> Hi Tomeu,
> On 10 Jun 2009, at 10:18, Tomeu Vizoso wrote:
>> On Wed, Jun 10, 2009 at 03:16, Andrés Ambrois<andresambrois at gmail.com>
>>> On Tuesday 09 June 2009 10:07:19 pm Martin Dengler wrote:
>>>> On Wed, Jun 10, 2009 at 01:46:07AM +0100, Gary C Martin wrote:
>>>>> There's another 'non-mesh' option possible for the '3 kids under a
>>>>> case. They will likely have no AP/infrastructure to connect to, so
>>>>> blanking the jabber server is not an option. For them the existing
>>>>> local-link config would be very viable solution, but there's no Sugar
>>>>> for it yet.
>>>> Would this be as simple as a "ad-hoc" device in the frame that put the
>>>> wireless into ad-hoc mode?
>>> As usual, Tomeu beat us all there :P
>> This is already in git HEAD, but the patch should apply cleanly to any
>> 0.84 such as Soas. If people would like to give it a try and provide
>> feedback, would be awesome.
> Beep. Just to note my interest on this...
> If we chose a fixed network name, rather than based on buddy name (much like
> olpc-mesh was), how would this behave when the actual network owner/creator
> shutdown while others were connected? I'm guessing at best it would trigger
> NM to recycle through the possible network connections again, hopefully
> landing back on the same (or re-creating if you happen to be first/only)
> named network.
I would expect that even if the initiator dies, the other laptops stay
in the ad-hoc network. I expected that the initiator is just the
initiator and has no special role once the first laptop joins the
network, but I may be wrong.
My only goal was to make work the specific scenario we tested in Paris
because I knew it would work up to that extent. If we can extend this
to work in more useful scenarios, would be great.
What IMO we shouldn't do is to promise that with our new, sweet
network technology we are going to solve all the networking needs
without need of infrastructure. Also would like to keep it really
simple from the user POV.
> Any one know if existing collaboration sessions would manage to stay up, or
> all die during this?
I think we can expect them to die whenever IP addresses and masks
change (just guessing again).
> I guess this needs testing... so if I have time will explore your patch with
> the 3 XO's here (sugar-jhbuild is not playing well for me just now with
> salute for testing this).
Awesome, very appreciated. I guess that as part of communicating this
new feature we would need a wiki page explaining its limitations.
>>> Sugar-devel mailing list
>>> Sugar-devel at lists.sugarlabs.org
>> Sugar-devel mailing list
>> Sugar-devel at lists.sugarlabs.org
More information about the Sugar-devel