The IPACS-project (Integrated Performance Analysis of Computer Systems) was founded by the Federal Department of Education, Science, Research and Technology (BMBF) in the program High-Performance-Computing to define a new standard for measuring system performance. One part of this research project is the design of a distributed architecture for execution of benchmarks on High Performance Computers. Its objective is to guide benchmarkers along a mostly automated benchmarking process cycle in compiling and executing benchmarks, and finally gathering and presenting the measured results on a central website. In this paper we present a distributed software architecture which allows immediate analysis of the results, is robust and flexible to support this benchmarking process for the benchmark community.