Re: [homenet] Let's make in-home ULA presence a MUST !?

Mikael Abrahamsson <> Thu, 16 October 2014 07:14 UTC

Return-Path: <>
Received: from localhost ( []) by (Postfix) with ESMTP id 4D5011A037A for <>; Thu, 16 Oct 2014 00:14:45 -0700 (PDT)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -2.967
X-Spam-Status: No, score=-2.967 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, HELO_EQ_SE=0.35, PLING_QUERY=0.994, RCVD_IN_DNSWL_MED=-2.3, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id flGuIT0Hd_cK for <>; Thu, 16 Oct 2014 00:14:43 -0700 (PDT)
Received: from ( []) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by (Postfix) with ESMTPS id 31FB11A0381 for <>; Thu, 16 Oct 2014 00:14:43 -0700 (PDT)
Received: by (Postfix, from userid 501) id 0074CA1; Thu, 16 Oct 2014 09:14:40 +0200 (CEST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple;; s=mail; t=1413443681; bh=su/lYHjI1bc1Vx8rOdF1DsWaqiKLAU2gEck6uLzyWYg=; h=Date:From:To:cc:Subject:In-Reply-To:References:From; b=WK80YcbSoQQ4/iVxt1VBpZj36gTEqwsqLkzdL01qQlJftBgZK44h22hRhay1Z8aL3 sUCHBR3KZSDIcKz9oc/aZF1n4zlWcztl6/5nQYZQvuvvR6U7RX207sytGYTDewnKMM g/G4l345FTJOU2mX8G8bBOBNm/yb8gJJizUhjNpA=
Received: from localhost (localhost []) by (Postfix) with ESMTP id ECCD09F; Thu, 16 Oct 2014 09:14:40 +0200 (CEST)
Date: Thu, 16 Oct 2014 09:14:40 +0200 (CEST)
From: Mikael Abrahamsson <>
To: Ted Lemon <>
In-Reply-To: <>
Message-ID: <>
References: <> <> <> <20141014142746.GX31092@Space.Net> <> <20141014145930.GY31092@Space.Net> <> <> <> <> <> <>
User-Agent: Alpine 2.02 (DEB 1266 2009-07-14)
Organization: People's Front Against WWW
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Cc: Erik Kline <>, HOMENET Working Group <>, Markus Stenberg <>, Pierre Pfister <>, Gert Doering <>
Subject: Re: [homenet] Let's make in-home ULA presence a MUST !?
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Thu, 16 Oct 2014 07:14:45 -0000

On Wed, 15 Oct 2014, Ted Lemon wrote:

> minutes is too short, because that will break for example a movie 
> stream.  But a 30-60 minute overlap will work fine for web browsing and 
> nearly all applications a typical end-user uses other than long movies.

I'd say a lot of the "movie" applications will work fine as well because 
a lot of them isn't a single TCP session over long time anyway, but that's 
another story.

What I most worry about is for instance SSH. However, I would like all 
applications to behave like "mosh" does (from a user point of view, not 
necessarily from a protocol point of view), in that an existing session 
can be moved between IPs, regardless if the change is gracefully handed 
over, or "hard" handed over (as in IP goes away, you get different 
connectivity with different IP 2 minutes later, and you can continue). We 
need this for video conferencing, VoIP calls, terminal sessions, basically 
*everything* that lasts more than 10-20 seconds on average.

This is of course not something the average application developer should 
need to handle, they should only need to use an API that says "please keep 
me connected to this DNS name" and then everything should be handled for 

> There are applications that will fail because they bind to an address 
> and have no error recovery mechanism for when the address goes away, but 
> AFAIK these are quite uncommon.  So really you are proposing a very 
> expensive cure for an insignificant problem.

I don't consider the problem insignificant.

Mikael Abrahamsson    email: