NTAC face-to-face
4/28/09
Attendees: Jeff Bartig, Andrea Blome, Sue Boff, Jeff Boote, Eric Boyd, Cort Buffington, Rich Carlson, Pat Christian, David Crowe, Cas d'Angelo, James Deaton, Dale Finkelson, Carla Hunt, Michael Lambert, Jen Leasure, Paul Love, John Moore, Dave Pokorney, Dave Reese, Chris Robb, Linda Roos, Chet Ruszczyk, Paul Schopis, Joe St Sauver, Martin Swany, Steve Wallace, Linda Winkler, Matt Zekauskus
- Call for Vice Chair nominations and Election of Vice Chair
o Vice Chair will take office immediately and will become chair on January 1, 2010
o John Moore nominated and accepted; none opposed.
- Future Research issues and next generation networking
o What should the next Internet2 network look like?
o Understanding the network design space.
Q: What is the next step in defining the process?
A: Try to identify areas to consider in design then understand the open questions to debate. Will start with initial brainstorming; then create subteams; then assemble the process. Would be helpful to identify the goal of the next network.
- TR/CPS Update
Steven Wallace and Dave Reese provided an update:
o Overall approach is a joint mutual working relationship. Going very well. Discussed tasks and details today.
o Having an agreement in place is necessary for the consolidated service.
o Have a good sense of requirements for he fall semester; have graphs.
o CPS needs upgrades: New connection to Ashburn 20G and upgrading NY from 20-30G; Chicago from 20-30G; Seattle from 2-10G.
o Working toward a consolidated service by August.
o Working together to ensure resources are consistent with a consolidated infrastructure.
o Have a very tight timeframe. Working with NTAC peering group to provide input and proceed quickly.
o Equipment? CPS has inexpensive HP switches in the exchange points. We are looking at upgrading to layer 3 devices and want to have the peer group input and be able to move along more quickly.
Q: Map - Capacity into exchanges; do we have adequate capacity to public exchanges? Moving to other consolidations, do we have sufficient peering at various locations with providers we have?
A: In the case of NY: 20G into exchange and 10G in top of public peering - the 10G is full so we're talking about adding another 10G. Chicago has the same situation 20G into Equinix - 10G into public is not enough and we will be looking at public ports. In Seattle, there is a 1g and an OC12 to another private peering - being upgraded from 6G to 10G. In Ashburn there will be one 10G public and one 10G for private peering. Ashburn changes the situation. Important peers that where individual peer connections were a big deal and will anticipate a 4thto Ashburn. Are currently looking at the data. GPN, Oregon and UEN put load on links in Kansas - Chicago to Kansas City and Kansas City to Salt Lake City. We may do something different here when we upgrade Seattle.
- IPv6/ARIN Update (Finkelson)
o Discussions on document sent to NTAC from the IPv6 working group. Presentation of that document generated quite a bit of conversation. Trying to understand that conversation. Highlights of document:
§ Technical concerns on readiness of IPv6.
§ Basically general agreement that it doesn't have all the functions but no reason to think that those can't be created. Tools have not yet been created or designed to help with those issues.
Concerns regarding the document:
§ First concern: In a practical sense would like to see a commitment of 100+ universities. This is a goal but not the primary purpose of the document sent out.
§ Second concern: Whether or not the approach suggested is appropriate within the activities of Internet2. It is not really up to this organization to make suggestions to campuses on what they should or should not be doing. There are a number of sections in the strategic plan that indicate these are discussions that are part of the Internet2 mission to at the very least, make available to the constituency the tools to build effective E2E communications within the community. May be seen as an outside agency coming in and may impact some people negatively. Far more likely to impact those who don't have the resources to do this on their own.
§ Third concern: What are we actually asking the NTAC to do? Not clear about the actual outcome envisioned when sending he document to the NTAC.
§ (Document highlights and NTAC recommendations were read)
o What trying to do with the document is have the NTAC say, 'may not agree with everything, but general perspective is something that this community needs to be made aware of.' We need to go through the various channels so this gets to the Board and they can instruct Internet2 staff to go out and build awareness.
o Needs to get to CIO's, Chancellors, etc. so they can go back to their institutions and make choices.
o Vote that NTAC will support delivering document to AOAC and adopted as a policy statement by Internet2.
o This is aligned with strategic plan implementation group who owns related tasks and this needs to get to them.
o It may be that the tasks are owned by the AOAC and if you look at the tasks to see where that fits, you can go there.
o It is not that Internet2 undertake to do this work but that they undertake to help to communicate this issue and provide information necessary to make a choice.
o This is a task that someone has to own and resources have to be allocated, etc.
o Will go to NTAC first, then to AOAC.
o This may already be Strategy #2; standards based tools and services.
o Is there a consensus that the views identified in the paper are reasonable ones that Internet2 should adopt as policy and promulgate to the community? Vote: None opposed, two abstentions.
Q: Any discussions on negative impact?
A: Last time folks were concerned about other subjects in relation to this i.e., concerned that this will be an unfunded mandate for campuses.
- Report from DCN Working Groups (Linda Winkler)
o Since last fall the Working Group has tried to grasp the concept of a fee recovery model; work concluded in March with recommendation to NTAC that fees be deferred to July of 2010.
o Have been working since February with Internet2 on DCN pilot service definition driven by AOAC to get service out of R&D and closer to a real product.
o Concerns:
§ Definition of cost model - don't have a sense to sustain this activity.
§ Relation of DCN cost to Internet2 network service portfolio - how does it relate to the rest of the network cost?
§ Question of value proposition for doing this - what's in it for the regionals and campuses to engage in new technology that does not understand the cost model?
§ While the original document discussed positioning backbone service, the community is really interested in E2E service. The regionals and campuses have to decide if they are going to step up to the same level of service and what is the end result.
§ This has all left us with more questions than answers.
§ At AOAC, CIOs were very concerned about cost and said to get a cost model out by the end of this year. We need to get a model by December so they can get it into their budgets.
- Update from Internet2 (Chris Robb)
o Network infrastructure Update
o New backbone links
§ Washington-Atlanta
§ KS - SLC
o Future planning
§ End of Life IP network equipment
§ Short-term effort to allow long-term architecture discussion to progress.
- Update from Peering Working Group (Jeff Bartig)
o Anything Internet2 purchases now will have to fit into a future converged consolidate network.
o Are discussing types of equipment to be used by peering service in the future.
o Peering working group will look at options considered and provide feedback to the process on concerns or issues of what's being proposed.
o Very tight timeframe.
- Miscellaneous
o May 5 NTAC meeting will be canceled.
o Send any reports to Linda Roos and she will forward.