The fermilab central computing facility architectural model

Abstract The goal of the current Central Computing Upgrade at Fermilab is to create a computing enviroment that maximizes total productivity, particularly for high energy physics analysis. The Computing Department and the Next Computer Acquisition Committee decided upon a model which includes five components: an interactive front-end, a LargeScale Scientific Computer (LSSC, a mainframe computing engine), a microprocessor farm system, a file server, and workstations. With the exception of the file server, all segments of this model are currently in production: a VAX/VMS cluster interactive front-end, an Amdahl VM computing engine, ACP farms, and (primary) VMS workstations. This paper will discuss the implemetation of the Fermilab Central Computing Facility Architectural Model. Implications for Code Management in such a heterogeneous environment, including issues such as modularity and centrality, will be considered. Special emphasis will be placed on connectivity and communications between the front-end, LSSC, and workstations, as practiced at Fermilab.