home *** CD-ROM | disk | FTP | other *** search
-
-
- IETF JOMAAN WORKING GROUP
-
- minutes of 2/6/89 Philip Almquist, secretary
-
- I: Announcement of User to User Connectivity Problems Working Group
-
- - new IETF group, chaired by Dan Long of BBN
-
- - will produce a paper on how end users can/should seek to resolve
- Internet connectivity problems
-
- - relationship to JOMAAN to be determined
-
- - unfortunately, neither group can fix broken hosts or broken
- administrators
-
- II: Mailing addresses
-
- - mailing list: njm@merit.edu (njm-request@merit.edu)
-
- - chairperson: hastings@psc.edu
-
- III: Charter
-
- - Gene will mail it out again
-
- IV: SNMP community names
-
- - all routers should support "monitor"
-
- - routers under the sole control of the regional NOC should support the
- NSFNET backbone community name
-
- - if neither of the above work to contact some gateway, try "public"
-
- - NSI "agrees in principle" to support community names that they will
- make available to regional NOC's
-
- - ditto for ESNET
-
- - regular polling of routers belonging to other organizations is a
- no-no, except that routers connecting two routing domains may be
- monitored by both NOC's (and should probably send traps to both NOC's).
-
- - the above restrictions on "regular polling" do not preclude sending
- queries to any router while actively debugging a problem
-
- V: Network maps
-
-
-
- 1
-
-
-
-
-
-
- - Merit is 90regional maps which are accessible via anonymous FTP
-
- - regionals which have maps available via anonymous FTP should send
- pointers to them to the njm list; Merit will treat this as an implicit
- request to regularly retrieve copies of the map
-
- - all maps should include a creation date
-
- VI: NSFNET <--> BBN core interactions
-
- - MERIT and DCA have been working on coordinating responses to
- mailbridge problems at the FIX locations
-
- VII: BITNET II
-
- - Scott Brim expressed concern that BITNET II is being designed by
- people who do not understand the Internet topology. Thus, the
- substantial new load it will place on the Internet may occur in
- inappropriate places. Scott will investigate further.
-
- VIII: Traceroute
-
- - several reported that third party traceroute is a real win, and hoped
- that other routers would support it soon
-
- IX: Appropriate us of the "status-reports" mailing list
-
- - the list is appropriate only for reports of current or very recent
- events, such as
-
- * "X will be down from ___ until ___"
-
- * "X is down"
-
- * "X was be down from ___ until ___"
-
- - Summary data can be interesting, but should be posted elsewhere
-
- X: FARNET Report (by Guy Almes)
-
- - FARNET wants increased FARNET<->IETF cooperation. Regionals should
- send people to IETF meetings; these people should report back to the
- regional operators and planners
-
- - periodic reports of usage/uptimes/etc. are useful (eg, the NSFNET and
- CERFNET monthly reports). People interested in helping to devise common
- reporting measures should send mail to Guy.
-
- - is application throughput commensurate with theoretical path
- bandwidths (ie, is performance as good as it ought to be)? This is an
-
- 2
-
-
-
-
-
-
- important question for assessing whether we run networks well and for
- justifying expensive, high-speed paths. Can we develop a "Dow Jones"
- average of network performance? Would this measure anything useful, or
- are most problems just broken TCP's that we have no control over?
- Interested parties should contact Guy about starting a joint
- IETF<->FARNET project in this area.
-
- XI: NSFNET information files
-
- - there was a request to the NSF NIC to provide a file of responsible
- persons indexed by network number
-
- - other ideas for similar useful files should be sent to nsfnet-info
-
- XII: NREN planning
-
- - Steve Goldstein of NSF wants input on how NIC's and NOC's should be
- organized in the NREN
-
- - Gene will send his ideas to the njm list; others may respond
-
- XIII: Whois service
-
- - NREN will use an X.500-based whois equivalent
-
- - some suggested that (in the shorter term) the existing NIC whois
- should be replicated on additional machines (this may not be practical)
-
-
-
- 3
-