Agree with the base requirements list: non-expert admin, non-expert
config, supports mesh networking of widely ranging link types, IPv4 and
IPv6 capable (optional: expandable for disseminating other configuration
information), existing protocol, multiple implementations, inter-op
demonstrated in production, runs on today's devices.
Disagree with jumping immediately to the conclusion that OSPF is
unsuitable to fulfill these requirements.
I've see plenty of super-simple OSPF networks which consist of 4 devices
communicating over an organizational boundary.
It's just a question of standardizing configuration and limiting options
to bare-bones e.g. default is that everything runs in area 0 => max of
±25 routers in the home = no big deal IMHO. It also leaves the future
open if there are people out there who do want to expand further e.g.
common building networks covering multiple residents.
Not that I'm pushing OSPF, but let's not prejudice the outcome before we
even start.
best regards,
RayH
Subject:
Re: [homenet] Question for you
From:
Randy Turner <[email protected]>
Date:
Mon, 3 Oct 2011 08:58:37 -0700
To:
Qiong <[email protected]>
CC:
"Pascal Thubert \(pthubert\)" <[email protected]>, C Chauvenet
<[email protected]>, MANET IETF <[email protected]>, Mark Townsley
<[email protected]>, JP Vasseur <[email protected]>, Acee Lindem
<[email protected]>, Fred Baker <[email protected]>,
"[email protected]" <[email protected]>, "[email protected]" <[email protected]>
Precedence:
list
MIME-Version:
1.0 (Apple Message framework v1084)
References:
<[email protected]>
<[email protected]>
<[email protected]>
<[email protected]>
<[email protected]>
<[email protected]>
<[email protected]>
<[email protected]>
<cah3bfadxemafskczjstknyz-6arrwgq9ccqs+ad4zxavz4a...@mail.gmail.com>
In-Reply-To:
<cah3bfadxemafskczjstknyz-6arrwgq9ccqs+ad4zxavz4a...@mail.gmail.com>
Message-ID:
<[email protected]>
Content-Type:
multipart/alternative; boundary=Apple-Mail-1-47901164
Message:
3
I would hope that we would NOT be seriously considering OSPF or IS-IS
in the home...this seems like using a sledgehammer to kill an ant.
How many routes are we talking about for a home network? I don't
believe any enterprise routing protocol was designed for a "zeroconf"
or "zeroadmin" type of environment. Our customers won't even know
what an IP address is.
Seems like a "RIP-like" (around the same scope of complexity) would be
enough for a homenet. I'm curious to see what comes out of the LLN
discussion.
The "filter" for any of these decisions should probably always be a
"zeroconf" or "zeroadmin" scenario -- if a proposed approach to a
problem can't exist in a "zeroconf/admin" environment, then I would
think it would not be the right choice. Also, as a "first cut"
solution, we I think we should be focused on the 80% use-case, not the
fringe. The participants of this working group, and their respective
home networking setups, are probably not our "typical" customer.
Randy
_______________________________________________
homenet mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/homenet
