--------------------
Thursday, 6 June 2013
Attending:
Dan Schmiedt - willys@clemson.edu
Michael Lambert - lambert@psc.edu
kevin mayeshiro - kmayeshiro@ucdavis.edu
akbar kara - akara@tx-learn.net
Rich Cropp - rac111@psu.edu
Agenda:
1) Agenda Bash
OpenDaylight update from Brent Salisbury.
* Big Switch has pulled out as a platinum sponsor of OpenDaylight.
* Hackfest goingon now, irc.freenode.net, #opendaylight channel. participation/observation encouraged even by non pprogrammers, inclusive general attitude.
* Big Switch L2
* OpenDaylight seems to be focusing on enterprise and is open source.
* Controller has a web interface, makes it accessible to network operators who are not programmers.
* Question from Steve W. i/r/t vendor response to CC-NIE. Brent highlighted operational use cases for SDN on a campus network.
Steve Wallace: ONF Plugfest going on this week, InCentre is the host. 50 engineers on-site. xx? vendors participating. Set up different switches using common controllers. Pretty much an all-day event, with very serious work and a lot of work for InCentre staff. Late July, expect to see "ONF-Stamped" certified Openflow-compliant products.
Steve discussed issues with OF spec as it relates to hardware. ONF is Working to put together a group of chip/hardware experts to bring parity to OF and hardware.
Steve went over Table-typing, TTP, where OF 1.3 tables are used to do things like control a router. They do this by exposing functional "knobs" via OF tables. For example, a FIB could be passed to a routing ASIC in this manner.
2) wg-SDN meeting at GEC17 (UWisc-Madison 7/21-23); NetGurus and wg-sdn meetings
There will be three events of specific importance to this community, in addition to all of the other stuff going at at GEC. Specifically:
a) wg-SDN meeting. During 1:30-3:30 slot on Monday, 7/22, will appear on agenda soon
b) SDN BOF dinner Monday evening
c) NetGuru, Tuesday afternoon, open invitation limited to first 35 people, nominally campus network architects but open to anyone.
3) New developments in OpenDaylight/BigSwitch
Big switch has pulled out as a "Platinum" member/board member/Tech Steering Committee member of OpenDaylight. Their exisiting contributions (code, etc) remain.
4) Intel ONP switch reference design http://semiaccurate.com/2013/04/17/intel-outs-two-new-open-networking-reference-designs/
Discussion. This is seen as a very positive thing and woudl really lend credibility to open networking, primarily beyond the giant network users, (Google, FB, Yahoo, etc) who can absorb massive amounts of risk. Now, it's easy to defend a white-lable switch purcahse, as "it's an Intel reference design."
5) IETF application and peering: ALTO (http://datatracker.ietf.org/wg/alto/ ) and CPNP (Connectivity Provisioning Negotiation Protocolhttps://datatracker.ietf.org/doc/draft-boucadair-connectivity-provisioning-protocol)
[SDNi which is related is at http://tools.ietf.org/id/draft-yin-sdn-sdni-00.txt ]
Discussion: this is generally seen as a good thing.
Jeff Handal mentioned that Steve Wallace's class on OpenFlow at NANOG was very good and well received. Significant improvment from the original class at Baton Rouge Joint Techs. Jeff asked Steve to clarify whether or not the ONF was puching a "standards based" approach. Steve answered that the ONF model is network-user driven, as opposed to the primarily vendor-driven (yet ostensibly open) IETF model.