Internet Info 1994 March/Internet Info CD-ROM (Walnut Creek) (March 1994).iso/inet/ietf/isis/isis-minutes-90july.txt

Text File | 1993-02-17 | 5.8 KB | 142 lines

CURRENT_MEETING_REPORT_



Reported by Ross Callon/DEC

IS-IS Minutes

The IS-IS Working Group met the morning of August 1, 1990, at the IETF
meeting in Vancouver, BC. We reviewed the most current Integrated IS-IS
specification.

The greatest amount of discussion was on the authentication field.
Several problems with the current text in the spec were pointed out.
Also, whatever we do will probably conflict with whatever the
authentication folks eventually tell us to do.  One option was therefore
to go back to what was originally in the spec, which is to leave the
contents of the authentication field unspecified.  However, there is an
urgent need for the most basic form of error supression.  For example,
it is very useful to provide a simple mechanism for preventing
mis-configuration of a single link from causing two large routing
domains to inadvertantly merge into one domain.

After a great deal of discussion, it was agreed that we would like to do
just about the same thing that OSPF already does:  provide a simple
password mechanism with an escape to allow future identification of
other mechanisms.  Ross Callon (as editor for the IS-IS specification)
was instructed to remove the details of the authentication field from
the main body of the spec, specifying the contents of the field as ``to
be determined'', and to provide an annex to the spec specifying how to
use the authentication field for carrying a simple password.  Also, we
agreed to use the same value for the authentication type field as used
by OSPF, in the off-chance that future assignments between
authentication type fields could be kept in alignment.

It was pointed out that the current definition of the manner of carrying
TAG information in the ``interdomain routing protocol information
field'' was difficult to process (in particular, it required that before
processing an ``IP External Reachability Information'' field, the
implementation would first have to check what the following field is,
and if it is an ``Interdomain Routing Protocol Information'' field, then
process the two fields in parallel).  After discussion, an alternate
encoding was agreed upon.

There was a discussion of the possibility that the amount of information
carried in the Inter-Domain Routing Protocol Information field may be
large, and that in some cases the bulk of level 2 routers (those that
don't do inter-domain routing directly) would therefore be required to
store information that they don't have any use for.  This would appear
to mean that folks determining how to use this field need to give
careful consideration to what inter-domain routing information should be
put into this field, and what should be carried by other means.  Ross
agreed to add a note to the spec describing this issue.

                                   1






The limit on the maximum number of addresses that can be assigned to a
single interface was discussed.  There was general agreement that
multiple IP addresses per interface was useful in some cases
(particularly for transition), but there was no obvious reason to limit
a router to two addresses per interface (as in the current spec).  It
was agreed that a better limit was whatever number of addresses could
fit into one occurrence of the ``IP Interface Address'' field in IS-IS
Hello packets, which implies a maximum of 63 IP addresses per interface.
It was agreed that this limit was plenty big enough, also that there was
no need to pick a smaller limit.

Rob Hagens pointed out that the use of the term ``segmentation'' in
section 3.6 was inconsistent with the terminology used in the OSI spec
(the meaning was consistent, just the terminology was different).  Ross
agreed to fix this.

It was agreed that after these changes were made, the spec was ready to
be published as an Internet Draft, and submitted as an RFC. Ross agreed
to send the draft spec to the Working Group first in case anyone could
find any nits.

A few other minor editorial nits were also transmitted to Ross during
side discussions.

Attendees


Karl Auerbach            [e-mail]
Fred Baker               [e-mail]
Art Berggreen            [e-mail]
Chet Birger              [e-mail]
Ross Callon              [e-mail]
C. Allan Cargille        [e-mail]
Curtis Cox               [e-mail]
Farokh Deboo             [e-mail]
Dino Farinacci           [e-mail]
Jeffrey Fitzgerald       [e-mail]
Chris Gunner             [e-mail]
Yong Guo                 [e-mail]
Robert Hagens            [e-mail]
Tony Hain                [e-mail]
Susan Hares              [e-mail]
Peter Harrison           [e-mail]
Kathleen Huber           [e-mail]
Paulina Knibbe           [e-mail]
Holly Knight             [e-mail]
Alex Koifman             [e-mail]
Gregory Lauer            [e-mail]
Walter Lazear            [e-mail]
Solomon Liou             [e-mail]
Yoni Malachi             [e-mail]
Douglas Montgomery       [e-mail]
Rebecca Nitzan           [e-mail]

                                   2






Zbigniew Opalka          [e-mail]
Brad Parker              [e-mail]
Michael Reilly           [e-mail]
Jim Reinstedler          [e-mail]
Jim Showalter            [e-mail]
Keith Sklower            [e-mail]
Frank Solensky           [e-mail]
John Veizades            [e-mail]
Chris Weider             [e-mail]
Steve Willis             [e-mail]
Walter Wimer             [e-mail]
Linda Winkler            [e-mail]
Allan Young              [e-mail]



                                   3