I'd suggest we basically recycle those retreat talks that cover
concrete results. I suggest we avoid promising to give talks on work we
plan to do (e.g., sting), but haven't yet completed (there's always next year).
We could have an entire session on internetworking:
Detour: The Case for Intelligent Internet Routing and Transport -- Stefan Savage
The Detour Packet Forwarding Engine -- Andy Collins
Modelling TCP Short Flow Performance -- Neal Cardwell
Towards Active Names: Framework and Applications -- Amit Aggarwal
thoughts?
tom