RELAY pursues research to achieve quantifiable improvements in resource utilization in distributed systems. We follow a bottom-up approach that aims at the development of mechanisms. We work on operating system changes, protocol extensions and new protocols, and advances in the architecture of distributed systems.

We want to extend the group on the following topics:

• Operating system support for the use of asymmetric multiprocessors. Examples of such asymmetric multiprocessors are found in systems using IBM/Sony's Cell processor, Intel's IXP network processor, or NVidia's newer CPUs that are supported by CUDA. These processors are currently used as an exclusive resource, and we want to change that.
• Mechanisms for next generation distribution systems. We pursue issues in operating systems, protocols and system architecture that will arise in the next generation of search and retrieval systems. We cooperate with international partners who will investigate searching, we will not investigate searching ourselves. We want to improve the infrastructure that is necessary to handle requests and to deliver the multimedia content that must be delivered to answer them.
• Operating system and protocol support for highly interactive distributed multimedia systems. Examples of such systems are multiplayer games, Internet telephony and conferencing, remote controlled systems and telesurgery. Reducing latency in highly interactive systems is a challenge that must be addressed at all levels: architecture, operating system, protocols, and application. We are looking for people with skills and ideas that fit the general field.

We may consider extending the group on the following topic:

• Media distribution systems. We feel that our work on non-interactive high-bandwidth delivery is well-staffed right now. For an applicant with an ideal background we may extend this activity anyway.

RELAY's research topics should result in research prototypes but we do not restrict ourselves to a particular research methodology. We apply experimentation with prototypes, simulation and analytical investigations as required by the research question.

We do not aim at qualitative improvements that may be achieved by new abstractions, middleware, or new developer tools. We may develop any of the above when it is necessary to achieve quantitative improvements.

People with a good insight in operating systems are frequently feeling intensely about their tools and platforms. We clarify therefore the following: In RELAY, we develop on Unix-like systems, the languages C and C++ are required but others are used as well, and we use latex for cooperative writing.

What application material should you include? Here is a list.