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

Text File | 1993-02-17 | 6KB | 146 lines

CURRENT_MEETING_REPORT_


Reported by Gene Hastings/PSC

NJM Minutes

There was discussion about the means to distribute timely up/down
reports.  Dale Johnson of Merit solicited comments on the value of
Network Status Reports, as the Merit mailings take extra effort.  The
general opinion is that they are still valuable but there is the hazard
of becoming inundated if everyone started posting campus up/down
messages.  Several results came of the discussion:


   o As an interim practice, operators continue as before, pending any
     observed instance of too much data.
   o Attempts will be made to use standard mail templates to issue
     messages from multiple places.
   o SURAnet will start posting messages about the status of FIX-East.
   o Participants will start to analyze and evaluate the possibilities
     of a database-like mechanism (X.500 was one such proposed) so that
     a network operator may query for status without needing to examine
     and classify messages for relevance simply to have context when the
     information is needed.



News, Comments:



   o X Drawing Tool:
     X based MacDraw work-alike:  idraw from Stanford U. Anon ftp
     Interviews.  Stanford.Edu) report courtesy Erik Sherk @ Cornell
   o Watch out for Byzantine Routing:
     Gene Hastings observed in November that connectivity is becoming
     rich enough with multiple mid-levels serving some regions that
     unexpected routing can result.  A specific example was a DuPont
     facility in Delaware whose path to the University of Delaware
     passed through at least four states:  DuPont  -> PSInet (Delaware
     -> Reston, Va.    -> Ithaca, NY  -> NSFNET  -> SURAnet (College
     Park, MD.)  -> Newark DE  -> U. of Delaware.
     There was once a time when such a path was considered extremely
     undesirable due to limited resources in the backbones.  In this
     example it may be moot, as the round trip times were fairly low.
     The major concern is that it is a likely harbinger of other

                                   1






     peculiar paths and that network operators must be aware of this
     situation in order to be productive when debugging.  Please Note:
     Sue Hares, as part of her examination of backbone routing changes
     would like to hear any reports of other unexpected paths; please
     send such reports to [e-mail].
   o Mac document translations:
     An issue of MacUser this past fall had an article and guide to
     document conversion tools, describing a number of utilities which
     can convert between different Macintosh graphics format.  In
     addition, there is a new version of DrawOver (3.0) released with
     Adobe Illustrator 3.0



New Business:



   o Ted Brunner - Map Edit:
     Ted gave an update on the map editor/topology database he and
     others have been working on at Bellcore.  This is an application
     and tools to create and store database entries on the behavior and
     configuration of a network (number and kind of interfaces, etc.).
     He showed the results of a prototype map editor which reads this
     database and draws a map based on that knowledge.  Copies of this
     software may be available to interested operators.  Versions exist
     for Sun 3, Sun 4, and DEC RISC. If interested call or mail to Ted.
   o Sue Hares - Routing Stability:
     Sue gave a presentation on route slop in the backbone, as seen by
     changes in the number of nets a given AS announces.  In some cases
     one could see individual nets toggling back and forth between
     alternate ASes announcing them to the backbone.  The first question
     raised is ``what does this mean''?  Is it affecting service to
     users?  This behavior is a general concern since frequent changes
     DO consume resources in the backbone and attached regionals.  In
     some cases it was possible to characterize a specific campus or
     net's activity as being a lousy line, with no desire or budget to
     correct it, or known itinerant service.  The implication of known
     sources of frequent routing updates raises the question of whether
     there is value in having ``pseudo-static'' routes, or the ability
     to set some hystereis on known sources of routing noise.



Attendees

Theodore Brunner         [e-mail]
Tom Easterday            [e-mail]
Fred Engel
Vince Fuller             [e-mail]

                                   2






Jack Hahn                [e-mail]
Susan Hares              [e-mail]
Eugene Hastings          [e-mail]
Dale Johnson             [e-mail]
Ken Jones                uunet!konkord!ksj
Dan Jordt                [e-mail]
Darren Kinley            [e-mail]
Walter Lazear            [e-mail]
Marilyn Martin           [e-mail]
Matt Mathis              [e-mail]
Bahaa Moukadam
David O'Leary            [e-mail]
Mark Oros                [e-mail]
Joel Replogle            [e-mail]
Robert Reschly           [e-mail]
Timothy Salo             [e-mail]
Tom Sandoski             [e-mail]
Bernhard Stockman        [e-mail]
Paul Tsuchiya            [e-mail]
Kannan Varadhan          [e-mail]
Carol Ward               [e-mail]
Linda Winkler            [e-mail]
Dan Wintringham          [e-mail]
Cathy Wittbrodt          [e-mail]
David Wood               [e-mail]
Sze-Ying Wuu             [e-mail]



                                   3