From OpenFlow Wiki

Meeting Minutes for SWAI, 09 April 2010

Future Meetings

• I'll be out on 4/22 and 4/29
• Planning on rescheduling 4/22 to 4/20.

Present

Glen, Tatsuya, Dan

Testing

Clearly I'm not using a big enough stick. Brandon has done some work and made some commits to the branch so we know that's working (with some changes). Thanks Brandon! I'll be working with Tatsuya and Glen to get a couple more tests developed soon.

Other Stuff

We discussed a number of other developments:

Atlanta Experimental Cluster

Met with Google about this this past week. It's taking shape. Goal is to have things up and running soon; specifically want to be able to submit papers with deadlines at the end of August (SigComm?) and in October (NSDI). Some discussion about whether LB4Gs can be used in place of LB6Bs for the aggregate/core.

HAL Definition

Initial definition of a hardware abstraction layer is under way, collaboration with Google and Nicira. Using Google protocol buffers for the current discussion.

NetFPGA

Some issues with Phys on 10G card. Tatsuya is finishing up on some of the SW work for gig card.

Testing is Exposing Flow Rate Issues

We should develop some flow table churn tests. Srini's work is showing up some issues with high flow rate demands.

Open vSwitch

We expect a transition to Open vSwitch as the primary switch platform going forward. The specifics of the transition haven't been determined yet. We should begin investigating this.