RSSAC Meeting, Minneapolis, MN
Minneapolis, Minnesota
RSSAC - 21st Meeting
Attendees:
| Present: | John Crain | Present via telephone: |
| Mark Kosters | Fredrico Neves | Daniel Karrenberg |
| Brian Coppela | Abram Thielke | Jessica Little |
| Nevil Brownlee | Johan Ihren | Mark Schleifer |
| Mike Green | Kenjura Cho | Kim Claffy |
| Jim Cassel | Jun Marui | Regrets: |
| Rex Hayes | Akira Kato | Gerry Sneeringer |
| Andrei Robachevsky | Yuji Sekia | Dave Swager |
| Steve Crocker | Lars Liman | |
| Geor Michaelson | Suzanne Woolf | |
| Cathy Murphy | Bill Manning | |
| Steve Conte | Joao Damas |
Agenda:
Root Operational Changes
- G. Anycast etc.
ICANN liaison report
ICANN nomcom
SSAC report
V6 status/issues
DNSSEC
Measurement
PR - web/presetnations/disclosure languages
publicity/press officer
Future
31jul2005 Paris
Agenda Correction: SSAC removed.
Root Changes:
G will move in topology & change administrative management.
It is moving
from the WDC area to Ohio and from SAIC to the US
DISA as the
operational team. Welcome to Mike and Jim from DISA.
A t the time of the topology move, the nodes will be upgraded. timeframe
for the physical move is estimated to be June-October. The Ohio node will
be physically much more secure than the old node. Will keep the same AS
if possible. May see AS path changethat removes AS 568 and injects AS
721.
Adding more processing power and bandwidth so "G" should be
seen as a better participant than in the past. Otherwise, externally,
there should be no visable changes.
D has delayed its
renumbering plan.
H has no current plans for restructuring.
Anycast Updates:
the root-servers.org public page ... it needs to be updated to
reflect WIDE/M links. Are there other updates required? Since
there is no common root-server operator organization, some
server operators are not keeping info on this page current. Once
such an organization exists, it will be prudent to have a common
site for publishing root server information.
ICANN liaison - Suzanne
had accepted the assignment. This is her report.
A few items worthy of mention:
- WSIS process & WGIG - a political effort - not a real root server issue ... we've done some good communications - need to do more.
- New TLDs - again not root ops.
- .NET rebid is underway.
Schedule calls for EOM-March to announce award - mid June for transition.
Still confusion over what we do. Process of root zone creation vs. server
operations.
Steve Crocker: Its been a good thing to have an RSSAC liasion and Suzanne Woolf is a good fit.
Suzanne Woolf: would like help in reviewing materials, etc.
Daniel Karrenberg: willing to help in my neighborhood. The ISOC briefings may be useful.
NOMCOM Suzanne Woolf:
we need to send a non-voting member. She did it last year and need to
send someone else. Lars Liman has been asked to volunteer but needs more
info.
Jun Murai: it seems a good idea to have someone other than Suzanne Woolf
take this on. Steve Crocker: almost more of a recruiting drive than a
vetting of a large candidate pool, Jun Murai might be the best choice.
Bill Manning: are we asking for a larger candidate pool?
Jun Murai: Yes, and will like to have an answer this week.
V6 Bill Manning:
S ummary of history - additional testing done; need to publish the results.
need to identify the communities impacted.
Issue: if an IMR
generates a priming query with at least one failure condition, it could
have severe impact on that caching server.
Jessica Little and Bill Manning did some testing. VeriSign did some work
and reported last the NANOG. Bill Manning will finish his testing, write
up the results and merge/compare results with the VeriSign data.
This will help identify the impacted IMR servers at that given
point in time.
Has potential for a pervasive and hard to detect, externally, negative impact on individual IMRs if turn on v6 glue for root servers.
Steve Crocker: Didn't
SSAC discuss this?
Fredrico Neaves: No, only discussed packet size issues and it was only
with regard to the impact on TLDs.
Johan Ihren: Discussed doing an inventory of old, broken software out
there that will get broken in new ways if add v6 to the root. What to
do about that broken software is a different matter. Know there is a problem,
don't know the scope or what to do about it?
Bill Manning: Before RSSAC can make a recommendation, should do research to identify what the problems will be. Will make ICANN more comfortable about when/how to proceed. Will try to compile info in time to review before the Luxemborg mtg. but only a portion of testing has been done. We need to compare data from each server instance against the failure modes
Steve Crocker: This is of critical importance.
DNSSEC Lars Liman:
deployment status - several facets. asking about
software status - there is much more - distribution et.al.
this is s/w only... update matches the last update as to timeline
A,J: 6-12 mos unless
pressure from last time
B: during first half of 2005, servers will run s/w
C: already has s/w
D: not repre
E: not repre
F: already running s/w
G: infrastructure upgrade and move takes priority; dnssec prep in planning?
no
H: runs diff s/w on diff servers; one already running s/w, other could
be made to within a week
I: less than 3 months if s/w is there and available
K: less than 6 months, cd be expedited if necessary
L: 6-9 months, cd be expedited if necessary
M: depends on performance; depends on release of BIND 9.4
Lars Liman: If want to publish that, also need to address ability to handle dnssec
Steve Crocker: active dnssec-deployment activity, would like to formalize this type of survey and work on next steps, add as part of the dnssec-deployment roadmap - online @ dnssec-deployment.org. Would it be possible to have responses by the MDP-ICANN meeting in a few weeks?
Room: doubtfull, too little time.
Steve Crocker: Would like to build positive pressure.
Note that the USG has selected ECC as approved algorithm. DNSSEC resolvers/validators will need to have this support added.
Bill Manning: what to do w/ old instances? upgrade? turn off? what? There is still significant traffic. They are no longer in the authoritative servers lists. Little said about upgrading systems to meet future needs when approval to renumber was given. This IP address is the last remaining from the original root servers. We have an obligation for stable operations. Bill Manning will make inquiries with USDoC for guidance. Options seem to fall into two general choices, shut down the server/service on the old address or continue to upgrade the service on the old addresses for undetermined periods. Further discussion should occur on the mailing list and a recommendation should be forthcoming.
Measurement:
cho. - not much progress. save Yuji has PhD... 5th WIDE/CAIDA next week.
Randy Bush was invited
to workshop at APNIC, and he had some comments about anycast DNS. Daniel
Karrenberg has not been able to replicate what he has observed, but it
could be because of where on the net it is happening. It's really mostly
about routing, not dns. Still get answers, but get
them from rapidly changing instances. There may be more data presented
at the CAIDA/WIDE workshop.
k.c. - OARC - next
rssac - rpt on 1918 @ root
dkf. - dnsmon data for the caida/wide mtg.
Public Relations
root ops vs rssac pages. clean up both. John Crain: SSAC has better structure
- can we learn from this? Steve Crocker: SSAC wants to produce better
reports. Looking for a "fellow" to track this. Considering some
candidates to track these efforts/changes
Jun Murai: Would like to promote what is happening, clear up misunderstandings;
more of an education role. Perhaps start with publication, starting with
the organized meetings.
The offical ICANN page does not have most RSSAC minutes i.e. Minutes are not being published.
Is a step just to publish minutes? Last time discussed, Bill Manning IS publishing raw minutes. ICANN was supposed to clean up these and post them.
Bill Manning knows someone who might be a good candidate to do it.
Suzanne Woolf would also like a contact person, someone to monitor things said about the roots. Roots don't have time to check for these, but if found, maybe could respond. This person can initiate contact, maybe dispute some low level findings.
Conclusion: So, yes, it would be helpful to have someone support the work of the group.
Jun Murai will write up a job description and send to the list for review. Once that is done, we'll work out how to fund that person.
Future
31jul2005 - Paris - 15:00-17:00
venue & room to be announced.
--scribes
Bill Manning
Cathy Murphy