Clustered Systems are usually designed for High Availability; continued service when component(s) fail. Some are special purpose/high performance designs. Here are some references for your reading on the subject: IMB HACMP high-performance cluster systems AlphaServer cluster with reduced risk