The IP over ATM Mailing List Archive by date

[asmith@Baynetworks.COM (Andrew Smith)]

Gerard,

Some other useful reading might be the current state of the
ATM Forum's LAN Emulation server-server protocols (95-1082r3) which
outlines a somewhat simpler architecture than either SCSP or Epidemic
which is targeted at somewhat more restricted topologies than what
people were indicating were desirable at IETF this week. I believe you 
already have access to this document and we intend to propose that ATMF 
make this generally available as an informational internet-draft as 
soon as possible in order for it to be considered as a 3rd alternative 
for NHRP/ATMARP, MARS and MPOA server synchronisation. 

It would seem to make sense to only have one protocol for all of this:
presumably MPOA would inherit the protocol that LANE chooses and, as one
of our customers pointed out to me today, the "M" in MPOA is supposed
to stand for multiple *data* rather than *control* protocols :-)

Andrew


********************************************************************************
Andrew Smith					TEL:	+1 408 764 1574
Technology Synergy Unit				FAX:	+1 408 988 5525
Bay Networks, Inc.				E-m:	asmith@baynetworks.com
Santa Clara, CA
********************************************************************************


> From owner-ip-atm@nexen.com Wed Mar  6 11:59:19 1996
> Date: Wed, 6 Mar 1996 11:42:40 -0800
> X-Sender: laubach@terra.com21.com
> To: "Gerard.GASTAUD" <Gerard.GASTAUD@vz.cit.alcatel.fr>, ip-atm@nexen.com
> From: laubach@terra.com21.com (Mark Laubach)
> Subject: Re: classical ip over atm
> Content-Length: 1276
> 
> Hi Gerard,
> 
> This is all changing a bit as a result of the rolc and ipatm meetings
> this week at the IETF.  There will be a separate document that will
> specify a single server synchronization protocol/system that will support
> the needs of  ATMARP, NHRP, and MARS.  The synchronization material will
> be removed from the classic2 document.  The classic2 document will then
> just continue to be just a bug fix of the original RFC1577 document. It
> will include pointers to the server synchronization work.
> 
> The server synchronization mechanism has yet to be specified.  Please
> regard the current material in the classic2-01 draft as "informational"
> at this time.
> 
> Mark
> 
> At 13:52 3/6/96, Gerard.GASTAUD wrote:
> >hi,
> >i am a novice in the subject so pls bare with me
> >1) we seem to have 2 proposals for server synchronization , one "base" on
> >epidemic database" present in 1577 and one from rolc namely scsp. it would
> >make sense to keep one, i guess the most efficient. Is there a policy to
> >make the rigth choice ?
> >2) 1577 feb 22 96 :
> >8.7.5.4origin address should be updated with received origin address and
> >not timestamp
> >8.7.6.1 & 2 forward reference to anti-entropy , there should be expalined
> >before appearing in text
> >8.7.2.8 & 8 could reference 8.4 instead of 8.3
> >gerard
> 
> 
>