Le 24/03/2015 17:49, David Lamparter a écrit :
On Tue, Mar 24, 2015 at 05:28:05PM -0500, Alexandru Petrescu wrote:
Le 24/03/2015 17:19, David Lamparter a écrit :
Before we lose this, let it be noted that we seemed to have
arrived at "no" for an answer to whether we want to deal with
non-transitive networks, *as part of this particular routing
protocol discussion*.
If I'm misrepresenting the outcome from today's meeting, someone
please correct me.
That transitiveness issue is from the fact that their routers only
have 1 interface.
I don't follow. If each of the routers in the A-B-C situation has a
wired segment attached to them, it's still A-C intransitive, but
they each have 2 interfaces?
Err, no. It's an A-B-C situation where each (even B) has 1 interface,
and all are in an IBSS. This is the situation described in that draft.
Are there 1-interface routers in homenet?
In an 802.11 IBSS, I would assume yes.
Well IBSS is not made to have 1-interface routers on them, so this
(1-interface routers) is not really good to do, in my personal oppinion.
IBSS is made to have 1-interface Hosts on it, not routers.
Remark this is my IMHO and a number of other people disagree with this.
I know.
I just tell that you can build a very good ad-hoc network without an AP
and with 2-interface routers. At that point there is no hidden-terminal
problem.
I *don't* think meshes are out of scope for homenet. I do think
meshes need a mesh routing protocol. But pulling this into the
current discussion seems to generate nothing but waste heat.
We dont have a definition of what a mesh is. Saying mesh is
inviting people from RoLL and MANET WGs to argue.
However, I'd doubt a homenet is a mesh in their sense.
Sorry - replace "mesh" with "network segment with intransitive
reachability" in my mail.
Ok. I'd substitute network for mesh altogether.
(And I'm not applying the concept to a homenet as a whole, I see it
as an attribute of a particular set of links / interface types.)
Well, I dont.
A mesh is a network and vice-versa.
As a consequence, when we talk about 802.11, we would be talking
about AP / BSS, not ad-hoc / IBSS.
Yes and no.
Yes, at home most deployments are in AP mode.
But no in that still at home the WiFi landscape has recently
become reacher than the old dichotomy AP-mode vs adhoc-mode. E.g.
the 802.11ac and ad products feature direct AP to AP communication
for range extension, or streaming from a tablet to a TV set, or a
LED projector switching between being an AP itself or being a
Client to another AP.
I'll have to look at these in detail, but they sound like individual
links that would be treated as P2P in the routing protocol.
I agree, although I dont understand what you mean by P2P.
Peer-to-peer networks is a great term used by DSL boxes and enhanced
Blueray players to download content from P2P platforms.
Peer-to-peer is also a great term used by the RPL protocol to tell a
node talks directly to another RPL node, instead of up-and-down a
Directed Acyclic Graph.
Point-to-point links are those which only have two IP ends on them, like
a cellular link of a smartphone, or the uplink of a DSL box.
A AP to AP wireless link could be a point-to-point link, but I dont
think point-to-point protocols are used on these links. The netgear
802.11ac AP to AP communication (see its user manual) is not necessarily
a point-to-point link. I _suppose_ that link is a WiFi link like any
other (i.e. not point-to-point link).
(In the tablet to TV set case, I guess routing wouldn't be involved
at all? Strays into the "is everything a router?" question,
though.)
Sure it should.
In a typical homenet you have a growing number of these apparently
specialized links. You have this tablet-to-TV but also 802.15.4 smart
light control, and remotely controlled window shades, alarms, smartbody
to smartphones, smartgrid control, and more. Each of these works ok
isolated in itself. When you want to link them together only IP works
for each. And only a notion of IP routing can find paths through such a
maze. Actually it's such complex that I gave up completely considering
it, it's craziness.
No hidden node problem. No intransitive reachability.
Massively reduced marginal links (because when you start losing
beacons between AP and Client, you'll be deassociated.)
Tinkering about this.
Tinker loudly, into your keyboard, onto a mail ;)
Overall it may boil down to have IP running on each of these links, a
routing protocol and an address and prefix assignment functionality.
Alex
-David
_______________________________________________
homenet mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/homenet
