Policy Statement on Address Space Allocations

Vadim Antonov avg at sprint.net
Tue Jan 30 02:44:42 UTC 1996

Well, i meant precisely the city loops which aren't multihomed,
and that's what customers are using to access telco facilities.

I.e. POTS customers (and their PBXes) aren't multihomed in the
sense used in the discussion.

The better term for telcos are doing for reliability between
switch sites is called "standby", which is *not* the same as
multihoming with adaptive routing.

The point was that high reliability may be achieved w/o tricky
things like adaptive routing,


>From sdenny at red.hex.net Mon Jan 29 21:14:22 1996
Received: from tiny.sprintlink.net (tiny.sprintlink.net []) by titan.sprintlink.net (8.6.9/8.6.9) with ESMTP id VAA09936 for <avg at titan.sprintlink.net>; Mon, 29 Jan 1996 21:14:21 -0500
Received: from red.hex.net (sdenny at red.hex.net []) by tiny.sprintlink.net (8.6.9/8.6.9) with ESMTP id VAA14645 for <avg at sprint.net>; Mon, 29 Jan 1996 21:14:11 -0500
Received: (from sdenny at localhost) by red.hex.net (8.6.10/8.6.10) id UAA02836; Mon, 29 Jan 1996 20:14:00 -0600
Message-Id: <199601300214.UAA02836 at red.hex.net>
From: sdenny at red.hex.net (Stephen Denny)
Date: Mon, 29 Jan 1996 20:14:00 -0600
In-Reply-To: Vadim Antonov <avg at sprint.net>
       "Re: Policy Statement on Address Space Allocations" (Jan 29,  4:19pm)
X-Mailer: Mail User's Shell (7.2.5 10/14/92)
To: Vadim Antonov <avg at sprint.net>
Subject: Re: Policy Statement on Address Space Allocations
Cc: nanog at merit.edu, sdenny at red.hex.net
Status: R

On Jan 29,  4:19pm, Vadim Antonov wrote:

> Note that regular telephone service does not generally employ multihoming
> and dynamic adaptive routing, but has a lot better reliability.  Something
> to think about.

}-- End of excerpt from Vadim Antonov

Regular telephone service is indeed generally multihomed except for the
final loop from the central office.

SS7 is a major pain partly because of the multiple routes it supports.
 - Multiple links in linksets (partly for capacity, partly for reliability)
 - Multiple linksets in routesets, where each linkset terminates on a
   different node

Typically the CO is "multihomed" to two mated STPs for signaling traffic,
and user traffic is routed through multiple tandems based on load,
availability, cost and destination.  Of course SCPs are also deployed in
pairs.  All this is in addition to the redundancy built into each piece
of equipment, or the underlying transport (e.g. sonet).  Of course none
of this stuff comes for the price of a Cisco router either.

What *is* true is that the telephone network is more or less a static
network in terms of routing (although becoming less so).  It would be
big news indeed if a new CO just popped onto the network and started
"announcing" itself!

This "static routing" idea makes for ultimate control and predictability
even with multiple routes.


Stephen Denny                        sdenny at hex.net
Member of the Technical Staff, Hex.Net Superhighway

More information about the NANOG mailing list