Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

[LinkSec] Teleconf notes 5/27/03




LinkSec Teleconf 5/27/03
Dolors Sala, chair, dolors@ieee.org
Allyn Romanow, notes, allyn@cisco.com

Attendees:
Antti Pietliainen, Dan Romascanu, Dolors Sala, David Nelson, Allyn
Romanow, Dennis Volpano, Tom Dineen, Mani Mahalingam


Summary
-------

Discussion of the 5 criteria as described in Dolors note on mailing
list 5/26
Minor disagreements and clarifications, see below.

Future work - update of 5 criteria. Will come to consensus at the
Interim meeting.

No teleconference next week, interim meeting in Ottawa, June 2-3.




==================================================
Discussion of attendance at upcoming meeting in Ottawa. If we have
trouble with attendance, we should see about co-locating with
802.3. This time, 802.3 was already meeting 3 weeks later than 802.1.

Discussion of 5 criteria, mail sent by Dolors

Broadmarket potential
Is it intentional to specify wireline exclusively?
While could be extended to larger scope, it is important to reflect current
interested parties. If we extend to a scope where people aren't
participating, we would not be able to deliver.
why use term public network?
Because Mick used in his scope.
Carriers would be a better name, we don't want to get into community networks
Re-word from "transition" to emergence of ethernet style connectivity into
subscriber and metro access networks
IF we mention RPR we need to mention metro access
It is important to describe only new work that needs to be done, not
work already undertaken


Conformity
802.1f - common definitions and procedures for 802 management information
a background chore that is always done

Distinct Identity
#2. discussion of "requires too many changes"
There is some disagreement whether 802.10 should be re-worked, however
there seems to be a consensus that we should start from scratch
.10 only needs tweaking, vs. .10 not enough on target to use it
We should have a slide for the presentation stating reasons why not using .10,
as this argument is likely to come up again

#3 Antti, 802.1 is bridging, whereas this, LinkSec, isn't bridging so it
shouldn't be an 802.1 document
802.1 is architecture as well
What we do within 802, has to be compatible with 802.1 bridging, an
historical requirement. Require interoperability with dot1 bridging
Just as dot1x is within the 802.1 umbrella, linksec also can be

Technical Feasibility - no discussion

Economic Feasibility
#1 - Similar technologies in 802.11 and IPSec, in both h/w and s/w.  Cost
factors known and in proportion. Have proven to be cost effective
solutions. The issue in a) is "known"

Discussion of key management being "out of scope" for the PAR.  The
consensus was that key management is indeed required as part of the
complete link security solution, that we would re-use existing, well
understood key management protocols as appropriate, and that the PAR
under discussion isn't the entire link security solution, but only the
Ethernet frame format portion.

So, yes, key management is "out of scope" for the current PAR but not
out of scope for the entire link security solution, which will likely
require follow-on PARs and projects.


Future work - update of 5 criteria

No teleconf next week as we have the Interim meeting