No root node
pete.cordell at BT-SYS.BT.CO.UK
Fri Sep 4 14:37:25 EDT 1998
In general, to a first order approximation, I would expect a fully
meshed arrangement of 'country' border elements. However I would see
the fine detail of the structure being down to who people have
arrangements for resolving addresses with. For example, BT might go to
AT+T to resolve all North and South American numbers, and AT+T might
come to BT for all European numbers. I could also see that
opportunistic companies could set up global clearing houses that would
handle address translation for outfits that haven't been able to
establish their own knowledge of all the borders such as ISPs, or
corporates. This would look like a root node to the people that used
it, but it would be used purely as a commercial expedient and not
mandated by the way the standard works.
On the issue of France moving its entry point to Lyon for the week, I
think we have to allow France to have more than one entry point. After
all, Paris could all of a sudden slide into the Seine, and the press, if
nobody else, would still want to make calls into France to find out
what's going on. By having multiple entry points, the French can turn
off their Paris BE for a while and switch on their Lyon one, and if the
data didn't propagate around the system for a while it wouldn't really
matter too much.
E-Mail: pete.cordell at bt-sys.bt.co.uk
Tel: +44 1473 646436
Fax: +44 1473 645499
>From: Santo Wiryaman[SMTP:swiryaman at VIDEOSERVER.COM]
>Sent: 04 September 1998 17:29
>To: ITU-SG16 at MAILBAG.INTEL.COM
>Subject: Re: No root node
>I am not at all opposed to eliminating the root domain. A couple of
>questions for Pete and the group:
>1. Are you suggesting that at the national level the border elements
>are fully meshed together, or alternatively arranged in a big ring as
>Radhika is suggesting?
>2. If so, when a new country "goes on-line" endpoints in the new
>country would not be reachable until the information is propagated in
>this fully meshed database (or inserted in the ring).
>3. Same if, say France, had to switch its National H.323 Border Element
>from Paris to Lyons for one week.
>4. Are you open to the idea of a global clearing house to resolve the
>address of border elements of countries which are currently not listed
>in the fully-meshed database.
>From: Pete Cordell [SMTP:pete.cordell at BT-SYS.BT.CO.UK]
>Sent: Thursday, September 03, 1998 10:20 AM
>To: ITU-SG16 at mailbag.cps.INTEL.COM
>Subject: No root node
>Apologies for not being on the last call, but...
>I feel one really important issue with Santo's contribution is that we
>must remove the root domain. Having such a concept would give
>politicians and lawyers far too much work!!! I believe the notes that
>put forward would allow the root node to be removed, and based on
>comments I heard from Jim, I believe what he is proposing would allow
>E-Mail: pete.cordell at bt-sys.bt.co.uk
>Tel: +44 1473 646436
>Fax: +44 1473 645499
More information about the sg16-avd