Offline computing farms

The D/null/ experiment will require substantial offline computing resources to perform production analysis, data reduction, and Monte Carlo event generation. Given the significant head start obtained by CDF, it is imperative that D/null/ have a fully functional offline computing environment ready at detector turn on to allow fast hardware and software shakedown and quick analysis of the data. The ability to get physics results out quickly is especially important to the graduate students, postdocs, and junior faculty who need to be able to present physics talks for future jobs/tenure. For these reasons, we believe D/null/ must begin planning its offline computing environment now, paying particular attention to the issue of software compatibility which has plagued CDF. In particular, proper evaluation of system cost should give at least equal weight to software development, management, and operational demands on the collaboration. We believe the best solution to the D/null/ offline computing needs is the development of computing farms utilizing DEC VAX computers. This note describes a model for such farms that we think can be realized with today's technology.