____________

NetSimQ User Interfaces

____________

User Interfaces


We use a multi-page layout, and configure the user interfaces into four pages:
  1. The topology page: Users may graphically specify a network topology by adding/deleting/moving nodes, adding/deleting links, and specifying both the link capacity at each link and the node buffer at each node.

  2. The network configurations page: Users may specify the traffic model, the routing algorithm/protocol, the resource reservation protocol, and the message scheduler, used in the simulation run. The choices currently supported in each module can be selected in a pull-down menu. The models/algorithms/protocols currently supported in NetSimQ are listed here. The parameters specific to a model/algorithm (e.g., the leaky bucket size and the token generation rate for the leaky bucket traffic model) are further inquired upon selection of a particular model/algorithm.

    Special care has been taken to ensure that all the inputs are ``consistent.'' For example, a PGPS scheduler has a known end-to-end delay bound only if it is used with the leaky bucket traffic model, and a stop-and-go framing strategy can only be used with the $(r,t)$-smooth traffic model. If users specify inconsistent inputs, a warning message is returned along with an advisory comment.

  3. The session configurations page: Users may choose to manually specify a session or randomly generate sessions based on certain probability distributions. In the former case, users are prompted for (i) the source(s) and destination(s) of the session; (ii) the start and end times of the session; (iii) the traffic characteristic of the session (using the traffic model specified in the network configurations page); (iv) the QoS requirements of the receiver(s) (e.g., the end-to-end delay bound and delay jitter bound); and (v) in the case of multicasts, when and in what sequence the group members join the multicast tree.

    In the latter case, users are asked to specify the probability distributions for generating (i) the number of sessions simulated, (ii) the call setup inter-arrival time, (iii) the time a session lasts, (iv) the traffic characteristic of a session, (v) the number of source/receiver nodes and the way to specify a node as a source or a receiver, and (vi) the QoS requirements of the receiver(s). Inputs to both the network configurations page and the session configurations page can be saved as configuration files for later retrieval.

  4. The run page: Users may start, suspend, or reset the simulation run. With the help of the monitor facility, users may request NetSimQ to provide instantaneous changes continuously or upon request. For example, users may view in a step-by-step manner how a multicast tree grows as all the group members join the multicast tree.

    Simulation statistics are categorized into three groups and displayed accordingly:

    • Global network performance: session rejection rate, current instantaneous network utilization, maximum network utilization ever observed, average network utilization, the ratio of injected traffic load to network capacity, average end-to-end delay, average delay jitter, average packet loss ratio, and the maximum queue length ever observed.

    • Session performance: maximum end-to-end delay which packets in a session experience, average end-to-end delay, average variance of end-to-end delay, delay jitter, packet loss rate, and the maximum queue length packets in a session experience as they travel from the source to the destination.

    • Node/link statistics: bandwidth (buffer space) currently available on a link (at a node) and instantaneous packet scheduling and transmission on a link.


Return to NetSimQ Overview Page
Date last modified -- August 3, 1998
Direct comments concerning this WWW site to: jhou@ece.osu.edu