Modeling And Simulation Technology - Internet Model Abstractions

  1. Discrete Models
    1. Queues
    2. Block Sizes (b) and Transfer times (t)
    3. Bandwidth -- bulk (max capacity) versus single link (transfer rate)
    4. Transission errors and packet loss (dropped)
  2. Multiple Sources (TCP, UDP, other)(wired vs wireless)
    1. Packet Collisions
    2. Retransmission Strategies:
      • TCP-Vegas
      • TCP-Reno
      • TCP-Newreno
      • TCP-FACK
      • TCP-SACK
      • many others
  3. Model Analysis Goals
    1. Steady State Behavior (for varius constant traffic rates)
    2. Transient Behavior
      1. Response to traffic spikes
      2. Response to changes of traffic levels.
    3. Using fluid flow approximation models for modeling congestion
      1. Linear differential equations
      2. Nonliner differential equations and Nonlinear "regions"
        1. Divergence and singularities
        2. Stable Points
        3. Attractors
        4. Transition regions
          1. Deterministic (time delay?)
          2. Non-deterministic choices (chaotic?)
  4. Robustness Analysis (survivability)
    1. Global "survivability" versus local needs satisfaction
    2. Techniques
      1. N-1 failure analysis
      2. N-m failure analysis
      3. Min cut/max flow
  5. Metrics
    1. Max network capacity?
    2. Point to point (P2P) network capacity
    3. Point to point (P2P) for multiple users
  6. Clusters, modularity, and reliability
    1. P2P in global network.
    2. P2P in local subnetwork (cluster module).
    3. Local resources, sharing excess.
    4. Local alternative for global resource requirements.
Updated: 2012-04-25