Message passing on a local network

This thesis focuses on understanding the cost of transparent interprocess communication in a distributed system consisting of a set of machines connected by a local network. Interprocess communication is transparent if processes can communicate without regard to physical host boundaries. Transparent interprocess communication is a very powerful tool because it allows us to view the collection of different machines as a single, logically unified computer system. We concentrate on the efficiency aspects of transparent interprocess communication on a local network. In order to obtain experimental evidence, a transparent message-passing mechanism has been implemented as part of the distributed V kernel. This message-passing mechanism has been used as the basis for various distributed applications. In particular, it has been used extensively for providing transparent file access from diskless workstations to a set of network-based file servers. Based in part on experience gained from the implementation and use of the distributed V kernel, this thesis presents four contributions: (1) An empirical evaluation of high-performance message passing on a local network. (2) A queueing network model of file access from diskless workstations over a local area network to a set of file servers. (3) An analysis of the protocol used to support the V interprocess communication on a broadcast network. (4) The integration of the broadcast and multicast capabilities of local area networks into message-based interprocess communication.

[1]  Willy Zwaenepoel,et al.  File access performance of diskless workstations , 1986, TOCS.

[2]  Peter A. Kirslis,et al.  A distributed UNIX system based on a virtual circuit switch , 1981, SOSP.

[3]  Robert H. Thomas,et al.  A resource sharing executive for the ARPANET , 1973, AFIPS National Computer Conference.

[4]  Kenneth P. Birman,et al.  A Local Network Based on the UNIX Operating System , 1982, IEEE Transactions on Software Engineering.

[5]  Willy Zwaenepoel,et al.  The distributed V kernel and its performance for diskless workstations , 1983, SOSP '83.

[6]  Paul V. Mockapetris.  Analysis of reliable multicast algorithms for local networks , 1983, SIGCOMM 1983.

[7]  Robert S. Fabry,et al.  A fast file system for UNIX , 1984, TOCS.

[8]  Maurice V. Wilkes,et al.  The Cambridge Model Distributed System , 1980, OPSR.

[9]  Elmar Holler.  The National Software Works (NSW) , 1980, Advanced Course: Distributed Systems.

[10]  Barbara Liskov,et al.  Guardians and Actions: Linguistic Support for Robust, Distributed Programs , 1983, TOPL.

[11]  D. W. Wall.  Mechanisms for broadcast and selective broadcast , 1980 .

[12]  James E. White.  A high-level framework for network-based resource sharing , 1976, AFIPS '76.

[13]  Keith A. Lantz,et al.  Structured Graphics for Distributed Systems , 1984, TOGS.

[14]  Barbara Liskov,et al.  Primitives for distributed computing , 1979, SOSP '79.

[15]  Andrew S. Tanenbaum,et al.  Distributed operating systems , 2009, CSUR.

[16]  Robert M. Metcalfe,et al.  Pup: An Internetwork Architecture , 1980, IEEE Trans. Commun..

[17]  Stephen S. Lavenberg,et al.  Mean-Value Analysis of Closed Multichain Queuing Networks , 1980, JACM.

[18]  Edward D. Lazowska,et al.  Quantitative system performance - computer system analysis using queueing network models , 1983, Int. CMG Conference.

[19]  Marvin H. Solomon,et al.  The Roscoe distributed operating system , 1979, SOSP '79.

[20]  A. J. Herbert.  The Cambridge Distributed Computing System , 1983, Advanced Course: Local Area Networks.

[21]  Robert Metcalfe,et al.  Reverse path forwarding of broadcast packets , 1978, CACM.

[22]  Elliott I. Organick,et al.  The multics system: an examination of its structure , 1972 .