next up previous contents
Next: network analysis Up: Applied network research: 1994 Previous: Contents


The Applied Network Research group agenda focuses on the evolution and future requirements of the data networking infrastructure. Research efforts include protocol performance evaluation for leading edge gigabit and ATM networks, traffic characterization for existing and future high speed networks, evolution and integration of higher level services and resources into the Internet fabric, and network engineering and planning support to the National Science Foundation.

In advanced network technologies, the ANR group participates in the CASA gigabit testbed project, specifically evaluating the performance of large-scale, distributed applications (including global climate change modeling) running across a heterogeneous supercomputer environment interconnected on a high-speed HIPPI based substrate.

In addition to investigation of experimental technologies infrastructure, ANR also participates in several collaborative research efforts with a perspective toward short to medium-term improvement of the existing operational infrastructure. In collaboration with other groups at SDSC, other NSF centers, and elsewhere, ANR pursues research in traffic and performance characterization to support the seamless integration of high end applications into the operational Internet. Because the nature of Internet workloads is changing so rapidly, ANR has undertaken studies on how operational Internet statistics collection should evolve with the changing workload in order to gain better insight into the increasingly dynamic nature of traffic. Also in support of the advanced functionality of the environment, ANR projects include issues in supporting servers for general data and information depositories, sophisticated user interfaces, collaborative and educational virtual environments, and high performance visualization of traffic characteristics.

ANR also participates in political developments in the Internet arena, working with NSF and other agencies in their role in advancing Internet evolution. The combination of the two perspectives, on advanced as well as existing network infrastructures, is in line with the ANR research agenda to develop synergy between political and technical objectives in Internet evolution.

This 1994 annual status report of the Applied Network Research (ANR) group at the San Diego Supercomputer Center (SDSC) and the University of California, San Diego (UCSD) represents an update on our research activities and outlines our future directions.

During 1994, the third year since the establishment of the Applied Network Research (ANR) Group, we have continued our work in the area of traffic characterization of strategic networking locations, in particular those which aggregate a large volume and range of traffic, typically NSFNET backbone nodes. An important component of our current research is the modeling of Internet traffic flows. In section 2, we discuss motivation for and potential applications of such an effort and describe recent findings. Topics of particular recent interest include traffic visualization, privacy concerns, metrics of burstiness, web traffic characterization, and caching and mirroring architectures for information resources.

We have and continue to apply the results of our investigations to operational and planned components of the Internet. Sections 4 discusses SDSC's involvement in the CASA gigabit testbed research project and in a multi-agency national ATM network project. Section 5 presents ANR efforts in NSF's NREN Engineering Group. In section 6 we discuss ANR's participation in the development, operation, and traffic characterization of network access points. In section 7 we discuss a method for dealing with privacy concerns on collected data.

In the coming year our agenda includes investigations of three Internet resource classes: information resources, communication resources, and computation resources. We discuss these areas in sections 8 through 10. In section 8 we discuss a proposed architecture for the establishment of a national hierarchical system for caching information resources. In section 9 we discuss our work on virtual collaborative educational projects, reflecting the second category, communication resources. In section 10 we discuss our planned projects in the third area, computation resources. Specifically we have proposed an infrastructure for supporting research on NSF's very high speed backbone network service (vBNS).

Section 11 describes the contributions made to the ANR research and analysis effort by three students from the Computer Science Department at the University of California, San Diego. Section 13 lists ANR publications released during the past year.

next up previous contents
Next: network analysis Up: Applied network research: 1994 Previous: Contents

Hans-Werner Braun
Wed Apr 19 20:12:08 PDT 1995