Various projects at OGI have been developing tools to adapt and specialize operating systems to meet various needs. The Synthetix project devised tools to specialize OS components, exploiting transient circumstances to enhance performance. The Quasar project is looking at adaptive resource management to provide multimedia Quality of Service in the presence of variable resource availablity. The Immunix project is investigating adaptation and specialization to ensure system survivability in the presence of security failures.

The products of these research projects are OS development tools, which are intended to be generally applicable to various operating systems. Generality has been demonstrated by applying them to radically different operating systems: HP-UX and Linux. Linux has been particularly valuable as an experimental platform, for the following reasons:

• extensive documentation about the system and its architecture
• freely available and disclosable system source code
• ANSI-C clean source code
• highly portable
• runs on low-cost and portable platforms, such as PC notebooks

Recent commercial pressure has lead us to consider applying these OS specialization tools to the Windows NT operating system. Our goal in attending this workshop is to learn about NT as a potential platform for applying our OS specialization tools. In particular, it will be enlightening to learn whether the assumptions that we have taken for granted in various UNIX systems are still relevant when working with NT.