In attendance:
David Klotz (DK) and Karl Levitt (KL)

TOPICS

1. New Items
1. KL: We have to contact Boeing about a no-cost extension, otherwise we have to do a lot of work and spend a lot of money by the fall
2. There is a PI meeting is 1 1/2 weeks, but our project is no administered under ITO instead of ISO, so we won't need to have a demo ready until August 2^nd. We should try to meet with Dan before that.
1. DK: I will be in Seattle on July 15^th, try to make the meeting for the 16^th.
3. KL: Did Jeff submit anything to the Applications conference? (No one was sure, but we doubted it).
4. KL: Are you still working on the cost model? DK: No.
5. Attacks on the FBI and the Senate
2. Nearest Neighbor Model
1. DK: Last week, Jeff and Jason worked on the Nearest Neighbor Model. Jason gave a demo/presentation at the Seclab Seminar
2. DK: Decentralized control - it may not be what Boeing wants. But the proofs (so far) don't get stuck in cycles like IDIP does.
1. DK: Jason is working on a minimal proof of correctness.
2. DK: Within one protection domain, you only need knowledge of your neighbors - you can pass information globally
3. KL: Similar to routing protocols? DK: Yes.
4. KL: Question of termination? DK: We think so.
1. KL: I imagine the experiments are harder than routing?
2. DK: There are bugs to be worked out, i.e. if there are 3 links coming off of a neighbor node or a double typology cycle.
5. KL: Does this fit into IDIP?
6. DK: Yes, we should present it as the next step of IDIP, an improvement on IDIP might fly better with Boeing. It has more stability, and the absence of loops.
7. KL: I don't think any else is doing a connection-oriented response (other than IDIP).
3. KL: For the PI meeting, we should try to do a demo
1. DK: There are standard simulators - NS (Network Simulator). We should be able to code it up quickly in Perl. Very simple and effective.
2. KL: We could do a simulation on a laptop. Chris Benkowin did a presentation off his laptop on block buffer overflow attacks.