Parallel I/o as a pArallel Application

Abstract

Metacomputing environments provide a software plat form for implementing parallel applications utilizing networked computing resources. Such programming environments offer a cost-effective alternative to paral lel machines for many classes of problems. Unfortu nately, most metacomputing environments either pro vide no I/O facilities, or serialize all I/O requests and thus limit application performance and scalability. PIOUS is a parallel file system designed to incorporate true parallel I/O into existing metacomputing environ ments. Because PIOUS is itself a parallel application, the functionality of a given metacomputing environ ment is extended without modification. This paper ex amines the PIOUS software architecture and program ming model and presents preliminary results from a prototype PIOUS implementation. Performance results are encouraging, demonstrating the potential of the PIOUS architecture for achieving scalable file system bandwidth in a network computing environment.

Get full access to this article

View all access options for this article.

References

Bernstein, Philip , Hadzilacos, Vassos , and Goodman, Nathan. 1987. Concurrency control and recovery in database systems. Reading, Massachusetts: Addison-Wesley.

Butler, Ralph , and Lusk, Ewing. 1992. User's guide to the p4 programming system. ANL-92/17. Mathematics and Computer Science Division, Argonne National Laboratory, Argonne, Illinois.

Cabrera, Luis-Felipe , and Long, Darrell D. E. 1991. Swift: Using distributed disk striping to provide high I/O data rates. Computing Systems 4(4):405— 436.

Cheung, A.L. , and Reeves, A.P. 1992. High performance computing on a cluster of workstations . In Proc. First International Symp. High Performance Distributed Computing, Syracuse, New York, September 1992.

Corbett, Peter F. , Feitelson, Dror G. , Prost, Jean-Pierre , and Baylor, Sandra J. 1993. Parallel access to files in the Vesta file system. In Proc. Supercomputing '93, pp. 472—481.

Crockett, Thomas W. 1989. File concepts for parallel I/O. In Proc. Supercomputing '89, pp. 574—579.

DeBenedictis, Erik , and Miguel del Rosario, Juan. 1992. nCUBE parallel I/O software . In Eleventh Annual IEEE International Phoenix Conference on Computers and Communications (IPCCC), pp. 0117-0124.

Dibble, Peter , Scott, Michael , and Ellis, Carla. 1988. Bridge: A high-performance file system for parallel processors . In Proc. Eighth International Conference on Distributed Computer Systems, pp. 154-161.

French, James C. , Pratt, Terrence W. , and Das, Mriganka. 1991. Performance measurement of a parallel input/output system for the Intel iPSC/2 hypercube. In Proc. 1991 ACM Sigmetrics Conference on Measurement and Modeling of Computer Systems , pp. 178-187.

10.

Hartman, John H. , and Ousterhout, John K. 1993. The Zebra striped network file system. In Proc. Fourteenth ACM Symposium on Operating Systems Principles, pp. 29-43. Kotz, David. 1993. Multiprocessor file system interfaces. In Proc. Second Intern. Conf. Parallel and Distributed Information Systems, pp. 194-201.

11.

Kotz, David , and Nieuwejaar, Nils. 1994. Dynamic file-access characteristics of a production parallel scientific workload. PCS-TR94-211. Dartmouth College. In Proc. Supercomputing '94 (to appear).

12.

Lamport, Leslie. 1979. How to make a multiprocessor computer that correctly executes multiprocess programs. IEEE Trans. Computers C-28(9): 690-691.

13.

Moyer, Steven A. , and Sunderam, V.S. 1994. PIOUS for PVM: User's Guide and Reference Manual (Version 1.1), 1994. Software and documentation available via the PIOUS home page located at http://www.mathcs.emory.edu/pious.html

14.

Patterson, David , Gibson, Garth , and Katz, Randy. 1988. A case for redundant arrays of inexpensive disks (RAID). In ACM SIGMOD Conf., pp. 109-116.

15.

Pierce, Paul. 1989. A concurrent file system for a highly parallel mass storage system. In Fourth Conf. Hypercube Concurrent Computers and Applications, pp. 155-160.

16.

Salem, Kenneth , and Garcia-Molina, Hector. 1986. Disk striping. In IEEE 1986 Conf. Data Engineering, pp. 336-342.

17.

Salmon, John. 1987. CUBIX: Programming hypercubes without programming hosts . In Proc. Second Conf. Hypercube Multiprocessors , pp. 3-9.

18.

Sunderam, V.S. 1990. PVM: A framework for parallel distributed computing . Concurrency: Practice and Experienee, 2(4): 315-339.

19.

Turcotte, L.H. 1993. A survey of software environments for exploiting networked computing resources. Technical report. Mississippi State University.