Modular Scalable I/O

Abstract We describe design issues for scalable I/O systems. I/O system modularity is the central issue. We identify the new information and algorithms responsible for high performance in scalable I/O. We show how to assign them to the parts of a scalable computer system so the result is modular. We have built a modular I/O system which is commercially available. We describe its design and performance benchmarks using it.

[1]  Hector Garcia-Molina,et al.  Disk striping , 1986, 1986 IEEE Second International Conference on Data Engineering.

[2]  G. Amdhal,et al.  Validity of the single processor approach to achieving large scale computing capabilities , 1967, AFIPS '67 (Spring).

[3]  Robert E. Benner,et al.  A radar simulation program for a 1024-processor hypercube , 1989, Proceedings of the 1989 ACM/IEEE Conference on Supercomputing (Supercomputing '89).

[4]  H KatzRandy,et al.  A case for redundant arrays of inexpensive disks (RAID) , 1988 .

[5]  G. C. Fox,et al.  Solving Problems on Concurrent Processors , 1988 .

[6]  Juan Miguel Del Rosario.  High Performance Parallel I/O on the nCUBE 2 (並列/分散処理論文 ) -- (ア-キテクチャ) , 1992 .

[7]  Randy H. Katz,et al.  A case for redundant arrays of inexpensive disks (RAID) , 1988, SIGMOD '88.

[8]  E. DeBenedictis,et al.  nCUBE's Parallel I/O with Unix Compatibility , 1991, The Sixth Distributed Memory Computing Conference, 1991. Proceedings.

[9]  E. DeBenedictis,et al.  nCUBE parallel I/O software , 1992, Eleventh Annual International Phoenix Conference on Computers and Communication [1992 Conference Proceedings].

[10]  Thomas W. Crockett,et al.  File concepts for parallel I/O , 1989, Proceedings of the 1989 ACM/IEEE Conference on Supercomputing (Supercomputing '89).