A Cluster is a parallel computing system constructed by interconnecting standard computers. The parallel performance of a cluster is often limited by performance compromises in conventional high-performance networks. This article discusses the benefits of a more specialized networking approach and the resulting design of a low-cost, high-performance interconnect optimized specifically to enhance both the parallel performance and high-availability aspects of a cluster. This interconnect, Memory Channel for PCI, attaches to the industry-standard PCI bus found in most computers. The first-generation product, running on a standard Digital UNIX Cluster, can provide cluster communication latency and overhead improvements of 100 and 1000 times respectively compared to conventional networks.