INTEL MEETING
June 2, 1999
2-3pm
1131 EU II
In attendance:

Karl Levitt (KL), Nicole Carlson (NC), Jim Hoagland (JH)
 

News
Negotiating Software Requirements
Brief Synopsis of Jim's Presentation
    1. News
      1. KL: I have just contacted Baiju about coming to visit in June or July.
        1. JH: I will be out of town June 25th-July 2nd. Back on July 6th.
      2. NC: I'm leaving the 18th, will likely have a job with NASA Ames until the Aug 6th when I come back for Summer Session.
      3. Next Week, we'll invite the GrIDS people to the Intel meeting - Chris, Jeff, Rick, Matt
    2. Negotiating Software Requirements
      1. KL: It a bit fuzzy
      2. Requirements Engineer and a Customer- how to extract what you really want. Consider performance requirements, function requirements, tradeoff
      3. JH: The customer may not know what they want.
      4. NC: Ask leading questions to determine what they want.
      5. JH: Price is a consideration
      6. KL: There may be conflicting requirements.
      7. Apply security requirements - come up with a model.
      8. JH: Negotiate security policy between a customer who needs security
        1. SCIPAD - Communication partner - what security you want
        2. Security Service Provider - medium size business negotiate with IBM for insurance, backups, port costs
        3. Exploration - software tools, plug in what you like, determine how it works with each system
      9. KL: Capture what each customer wants as a final state machine
    3. Brief Synopsis of Jim's and Nicole's Presentation
      1. JH: We've worked on 3 situations of transferring GrIDS into LaSCO
        1. Simple Policy - one edge in GrIDS
        2. One edge and additional state
        3. Two edges to trigger and external state
      2. JH: Problems with the isolated node. Add more to the rulesets, determine how to solve it.
        1. The way we maintain external state with separate graphs with information accumulating. Merge in with reports coming in. Reason why not sufficient for lone node policy.
        2. The order of the merged graphs matters, need to label them to distinguish them or add functioning
      3. JH: After we solve the isolated node problem, then we'll look at an edge and an isolated node, then hierarchy.
      4. NC: I looked at the GrIDS technical report, looked at the ruleset, but I don't know Perl yet.
      5. JH: You had to look at the code for the semantics.