作者: G. Hipper , D. Tavangarian
DOI: 10.1016/S1383-7621(97)00037-4
关键词: Communications protocol 、 Workstation 、 Network architecture 、 Interconnection 、 Component (UML) 、 Protocol (object-oriented programming) 、 Parallel computing 、 Cluster (physics) 、 Computer science 、 Domain (software engineering) 、 Computer architecture
摘要: Abstract In this contribution, recent approaches to increase the communication performance of workstation clusters for use as a multicomputer (i.e. parallel computing in clusters) are classified and described. Well-defined examples given demonstrating improvement through high-speed interconnection networks optimized protocol system architectures. A special interest within area is detailed description research domain cost-efficient with standard component systems, concentrating on Beowulf Parallel Workstation Concurrent Network Architecture (CNA) cluster.