Scott Bradner, senior technical consultant and director of the Harvard Network Device Test Lab at Harvard University, gave the keynote address at the QoS workshop. Photo by Tom Mead

Delivering QoS to NASA Applications

By Marjory J. Johnson

In the past NASA has supported critical network applications by dedicating separate communications circuits for individual missions, a very expensive approach that results in networks that lie idle much of the time. Emerging technological advances will enable more cost-efficient sharing of network resources by delivering "Quality of Service" (QoS) guarantees to selected user applications. QoS allows users to request special services, such as reserved bandwidth or maximum delay, in transmitting their information over a network. That is, QoS will allow designated applications to receive preferential treatment at the expense of lower priority applications when network resources are inadequate to satisfy all requests.

Because QoS is critical to NASA missions, the NASA Research and Education Network/Next Generation Internet (NREN/NGI) Project Office hosted a QoS workshop at NASA Ames Research Center on August 18 and 19. Attendees included NREN partners from other government agencies, universities, industry and the Internet2 community, which is a collaborative effort by more than 120 U.S. universities to coordinate the development, deployment, operation and technology transfer of network services to support advanced applications.

Different applications require different levels of network service. Typical QoS performance parameters include bandwidth, delay, jitter (random variation in the timing of a signal) and loss. In the workshop keynote address Scott Bradner, senior technical consultant and director of the Harvard Network Device Test Lab at Harvard University, described QoS as the ". . . unfair allocation of resources under congestion conditions." In other words, during periods of congestion, applications with QoS guarantees will receive preferential access to limited network resources.

Development of technology for providing QoS guarantees across wide-area networks is currently a major research effort in the networking community. Several mechanisms for delivering QoS guarantees are being developed, and various approaches for providing end-to-end QoS have been proposed. Since QoS is meaningful only in the context of specific application requirements, the workshop opened with a discussion of two specific NASA applications: digital video distribution to support shuttle-launch operations and Earth Observing System data distribution. The bulk of the workshop was spent discussing specific QoS research areas, including QoS routing and middleware (software that mediates between an application program and a network), experimental measurement of QoS deliverables, plans for implementing a QoS infrastructure for Internet2 and QoS policy.

There was general agreement at the workshop that a better understanding of application QoS requirements is necessary to develop an effective QoS network infrastructure. NREN activities to implement NASA revolutionary applications will make a major contribution toward this effort. There was also general agreement that QoS is a community problem, as an end-to-end application is likely to cross several network domains belonging to different network providers. NREN anticipates continuing discussion and collaboration with its networking partners to address QoS challenges for the Next Generation Internet.

For more information, see: http://www.nren.nasa.govworkshop_home.html