____________

Current Work

____________

I. The first software release of NetSimQ is scheduled in September 1998. We are currently (i) writing user's manual and developer's manual; (ii) testing the fidelity of NetSimQ by comparing the simulation results with real-world performance data obtained from the Ohio Computing and Communication Research Testbed (OCARNet); and (iii) integrating off-the-sheff graph plotting facility with NetSimQ to facilitate generation of two/three-dimensional plots and pi diagrams for simulation results.

We will also seek industry partners to lay a roadmap for commercializing NetSimQ.

II. We will continue to implement the proposed protocol stack in MK 7.3 using the CORDS system (which is itself an extension of x-kernel with the notion of path introduced). We expect to have an operational system running by the end of summer 1999.

We will also take advantage of the GateD Multicast source codes (distributed by the GateD Consortium, Merit Network, Inc.) and prototype the proposed QoS-driven core-based tree multicast routing protocol both as a stand alone mechanism and as part of the protocol stack in NetBSD.

III. We will develop a packet eligible time calculation mechanism along with its associated information update and buffer management strategies. The mechanism, coupled with any multicast routing protocol that provides the end-to-end delay bound, can provide QoS services to the following four categories of applications:
  • applications with the bounded end-to-end delay requirements;
  • applications with the bounded delay and inter-message delay jitter requirements;
  • applications with the bounded delay and inter-destination delay jitter requirements; and
  • applications with the bounded delay, inter-message delay jitter, and inter-destination delay jitter requirements.
    • IV. We will devise a mechanism to determine, given the requested-QoS region of a new connection, a suitable offered QoS in the requested-QoS region for the connection such that the QoS requirements of both the new and existing connections are fulfilled. The motivation for this work is that the QoS requirements in most of the QoS models available are specified by fixed values. This may be a simplistic approach because it does not take into account of the fact that for many applications the QoS requirements can be specified in terms of a region in the QoS parameter space. We will develop, based on the IETF-specified QoS service parameters, an adaptive QoS specification and management model that exploits the above fact.
    Return to Project Home Page
    Date last modified -- August 3, 1998
    Direct comments concerning this WWW site to: jhou@ece.osu.edu