NTAC 4/3/12 Notes
WG Updates:
IPv6: Not much to report. There has been some discussion around World IPv6 Launch Day. The IPv6 workshop is also being reworked.
Peering and Routing: The group worked through the request from the GRNOC to drop some messages from Dual homed connectors when there are outages, made some suggestions to the NOC.
TR-CPS upgrades at both Ashburn and Equinex. They are close to turning up the lambda across the southern route.
SDN: There are some ideas circulating on the list and conversations continue on work on the charter.
General Updates:
I2: The DC-Chicago 100G transport has been put in place. The SNMP polling issue has been resolved. The LAG in Chicago to support TR-CPS has been put in place.
AOAC:
The Architecture and Operations Advisory Council met on 14 March 2012. Dave Gift provided an update on the work of the AOAC fee subcommittee. The subcommittee focused first on the goals and principles for setting Internet2 network fees and began to review the current fee model as well as some alternate models. The subcommittee has requested additional information from Internet2 staff and will meet in April to continue their work. Charlie McMahon provided an update on the progress of the SIP RFP being reviewed by a committee of community members. Rob Vietzke provided an update on the Internet2 Network build from Seattle to Chicago. Discussions continue and solutions continue to be sought to provide that critical piece of the Internet2 infrastructure. The Council discussed the upcoming changes to the governance structure. The current Councils will be replaced by Policy Advisory Groups (PAGs) in conjunction with seating of the new Board this spring. More information will be forthcoming on these changes at the SMM. Finally, the Council continued their ongoing discussion of the Innovation Platform. The community review of the Innovation Platform equipment RFI was discussed as well as other progress on and plans for the Innovation Platform.
NDDI Discussion:
Chris Robb gave a summary: We recognize there are some process issues and that there has been no clear charge from the AOAC. He encouraged everyone to think about what they would like to see the AOAC issue as a directive to the NTAC. There has been some confusion, in general SDN is the future of what we are doing but we will not take risks with the network. The focus in this will be risk mitigation.
Grover: We are looking for a list of features and tools. There is a high focus on risk mitigation. Dual connectivity should be a goal for connectors. There are certainly some issues about security and the use of SDN.
John Moore: There is value in stepping back and taking a look at the process. Internet2 definitely needs to differentiate itself from standard carriers be out in front of the technology curve. Other areas that can happen in are Net+ and national distribution of those services.
Grover: Johns view matches with his. We want a platform where there are many ways to innovate while at the same time maintaining a production level of service.
Akbar: We anat our IP services to remain intact but we need to move the network forward. He suggested an approach which he detailed in a later email to the NTAC.
Dan Schmeidt: In talking to faculty it seems that the Open Flow infrastructure gives them a much greater chance to use all of the capabilities of the network, in a want that earlier implementations of circuits has not. Bill Owens pointed out that the NDDI infrastructure is there for that purpose.
Matt Davy: This is not an either/or situation. SDN is just a feature, that can be turned on in the boxes or not. We want to work with the vendors to develop this because that is the innovation path. SDN will be implemented in the backbone only when it is ready.
Grover: Pushed back on the notion of building a separate infrastructure. it tends to waste money, its hard for connectors and researchers to get at and use. Parallel networks are not as useful, OPEX is generally a problem and connecting to them is often a problem. Bill O. pointed out that this really goes both ways, some users really want a totally separate infrastructure. Matt Davy stated that a separate structure exists in GENI and that will always be in place for those users.
Grover: Make the backbone 100G first and then augment a platform that can be expanded into SDN as that is appropriate.
Michael Lambert asked about timeframes and how much this is being driven by BTOP. Chris Robb stated we are in fact being driven by the CIO community as well as by BTOP constraints. It would also be nice to be able to do this work over the summer from a logistical point of view.
Jeff Batig asked if there would be some sort of press release about the RFP at the SMM. The answer was we are not sure.
Akbar: How ands when does an architecture get developed? This is really the key set of choices that must be made. Bill Owens agreed that there needs to be a direction forward for the group to see what the architecture might look like. Chris indicated he would resend a note about what decisions have been made vs what areas are open.
Paul Schopis: He stated he did not think there was a great deal of flexibility remaining. SDN is the direction forward, NDDIOS3E is moving forward. There will be some areas where there is room for discussion but the base decisions seem made. Bill O. agreed saying that if those choices are made then the way forward is to advise how to move the network forward in the areas they really are interested in and care about.
Matt Davy stated that we need to not just innovate on the network we also need to innovate the network. He asked how else we could do this if we are not using SDN?
Chris laid out these points: The network needs to be stable, it needs to provide advanced services for those that want them, it needs to provide a platform where all of this can happen.
Akbar agreed that there are many competing issues here. He urged everyone to keep in mind that SDN will only be put into the network when it is ready.
Michael Lambert
Matt Davy
Akbar Kara
Steve Wallace
Eric Boyd
Jeff Schiller
John Moore
Grover Browning
David Crowe
Dan Scheimdt
Dave Pokorney
Shannon Spurling
David Wood
Shuman Huque
Bill Owens
Paul Schopis
Matt Zekauskas
Jeff Bartig
Brian Cashman
Ryan Vaughan
Tom Knoeller
Chris Robb