Summary
High performance distributed computing systems require high performance communication systems.F-channels andHierarchical F-channels address this need by permitting a high level of concurrency like non-FIFO channels while retaining the simplicity of FIFO channels critical to the design and proof of many distributed algorithms. In this paper, we present counter-based implementations for F-channels and Hierarchical F-channels using message augmentation-appending control information to a message. These implementations guarantee that no messages are unnecessarily delayed at the receiving end.
Similar content being viewed by others
References
Agrawal S, Ramaswamy R: Analysis of the resequencing delay for M/M/M systems. In: Proceedings of the SIGMETRICS Conference on Measurement and Modeling of Computer Systems. ACM 1987, pp 5–12
Ahuja M: Repeated global snapshots in asynchronous concurrent systems. Technical Report OSU-CISRC-8/89 TR40, Department of Computer and Information Science, The Ohio State University, August 1989
Ahuja M: Flush primitives for asynchronous distributed systems. Inf Process Lett 34(2): 5–12 (1990)
Ahuja M: Designing concurrent systems using hierarchy of communication speeds. Technical Report OSU-CISRC-1/91 TR1, Department of Computer and Information Science, The Ohio State University, January 1991. This research also appears in: Ahuja M, Prakash R, (eds) On relative speed of messages and hierarchical channels. In: Proceedings of the Symposium on Parallel and Distributed Processing, pp 246–253, 1992
Ahuja M: Assertions about past and future in highways: global flush broadcast and flush-vector-time. Inf Process Lett 48(1): 21–28 (1993)
Ahuja M: An implementation of F-channels. IEEE Trans Parallel Distrib Syst 4(6): 658–667 (1993)
Ahuja M: Faster communication using multiple and heterogeneous paths. Technical Report CS94-371, UCSD, May 1994
Ahuja M: Highway's position with respect to CATOCS. Technical Report CS94-372, UCSD, May 1994
Ahuja M: Receiving messages in receive-phases and incrementally publishing the needed partial order. Technical Report CS94-343, UCSD, August 1994
Ahuja M: Assertions about past and future: communication in a high-performance distributed system highways. In: Proceedings of the Symposium on Parallel and Distributed Processing. IEEE December 1993, pp 241–244
Ahuja M, Raynal M: An implementation of global flush primitives using counters. To appear in: Parall Process Lett 5(2), 1995
Ahuja M, Varadhan K, Sinha A: Flush message passing in communicating sequential processes. In: Rishe N, Navathe S, Tal D (eds) Parallel architectures. IEEE Computer Society Press 1991, pp 31–47
Birman K: A response to Cheriton and Skeen's criticism of causal and totally ordered communication. Oper Syst Rev January 1994
Birman K, Joseph T: Reliable communication in the presence of failures. ACM Trans Comput Syst 5(1): 47–76 (1987)
Birman K, Schiper A, Stephenson P: Lightweight causal and atomic group multicast. ACM Trans Comput Syst 9(3): 272–314 (1991)
Camp T, Kearns P, Ahuja M: Fast batched data transfer with flush channels. Technical Report WM-93-1, The College of William and Mary, 1993
Cheriton D, Skeen D: Understanding the limitations of causality and totally ordered communications. In: Proceedings of the Fourteenth Symposium on Operating Systems Principles. ACM, December, 1993
Chowdhury S: The mean resequencing delay forM/H k /∞ systems. IEEE Trans Softw Eng 15(12): 1633–1638 (1989)
Hoare CAR: Communicating sequential processes. Commun ACM 21(8): 666–677 (1978)
International Organization for Standardization. ISO Open System Interconnection — Basic Reference Model ISO/TC 97/SC 16N719, 1981
Kearns P, Camp T, Ahuja M: An implementation of Flush channels based on a verification methodology. In: Proceedings of the 11th IEEE International Conference on Distributed Computing Systems. IEEE, 1992, pp 336–343
Lai T, Wu L: An (N−1)-resilient algorithm for distributed termination detection. In: Proceedings of the Symposium on Parallel and Distributed Processing. IEEE, 1992, pp 274–281
Malki D, Amir Y, Dolev D, Kramer S: The transis approach to high availability cluster communication. TR 94-14, Inst of Comp Sci, The Hebrew University of Jerusalem, June 1994 (submitted for publication)
Peterson L, Buchholz N, Schlichting R: Preserving and using context information in interprocess communication. ACM Trans Comput Syst 7(3): 217–246 (1989)
Raynal M, Schiper A, Toueg S: The causal ordering abstraction and a simple way to implement it. Inf Process Lett 39(6):343–350 (1991)
Schlichting RD, Schneider FB: Using message passing for distributed programming: proof rule and disciplines. Trans Program Lang syst 6(3): 402–431 (1984)
Shafer K: Modeling, implementing, and tools for studying logical process-to-process channels in asynchronous distributed systems. Phd Thesis, The Ohio State University, March 1992
Shafer K, Ahuja M: Process-Channel agent -Process model of asynchronous distributed communication. In: Proceedings of the Twelfth IEEE International Conference on Distributed Computing System. IEEE, 1992, pp 4–11
Shafer K, Ahuja M: Programmer's manual to a C++ implementation of hierarchical F-Channels. Technical Report OSU-CISRC-6/92-TR-18, The Ohio State University, June 1992
Shafer K, Ahuja M: Distributed modeling and implementation of high performance communication architectures. In: Proceedings of the Thirteenth IEEE International Conference on Distributed Computing Systems. IEEE, 1993, pp 56–65
Tanenbaum A: Computer Networks (2nd Edition). Prentice Hall, Englewood Cliffs, NJ, 1988
Yum T-S, Ngai T-Y: Resequencing of messages in communication networks. IEEE Trans Commun COM-34(2): 143–149 (1986)
Author information
Authors and Affiliations
Additional information
Keith Shafer received the B.A. degree in computer science and mathematics in 1986 from Mount Vernon Nazarene College, Mount Vernon, Ohio, USA, and the M.S. and Ph.D. degrees in computer science from The Ohio State University, Columbus, Ohio, USA, in 1988 and 1992, respectively. He is currently a Senior Research Scientist at OCLC Online Computer Library Center Inco, Dublin, OH, USA. His research interests include tools for comparing logical channels and methods for automatically constructing corpus grammars from tagged documents as an aid for database preparation and document conversion. Dr. Shafer is a member of the IEEE Computer Society.
Mohan Ahuja received the M.A. degree in 1983 and the Ph.D. degree in 1985, both in computer science, from the University of Texas at Austin. He is currently with Department of Computer Science and Engineering, Univ. of California, San Diego. His recent research contributions include Global Flushing, message receipt in Receive-Phases, Incremental Publication of a Partial Order, Design of Highways (a high-performance distributed programming system) and — in collaboration with others — Passive-space and Time View, Performance evaluation of F-Channels, and Units of Computation in Fault-Tolerant Distributed Systems. His current research interests are in high-performance distributed communication and computing architectures, building high-performance systems, distributed operating systems, distributed algorithms, fault tolerance, and performance evaluation.
Parts of this paper appeared in two conference papers, (1) “Distributed Modeling and Implementation of High Performance Communication Architectures,” in proceedings of the Thirteenth IEEE International Conference on Distributed Computing Systems, papes 56–65, 1993 and (2) “Process-Channelagem-Process model of asynchronous distributed communication,” in proceedings of the Twelfth IEEE International Conference on Distributed Computing Systems, pages 4–11, 1992
Rights and permissions
About this article
Cite this article
Shafer, K., Ahuja, M. Implementation of hierarchical F-channels for high-performance distributed computing. Distrib Comput 8, 211–218 (1995). https://doi.org/10.1007/BF02242739
Received:
Accepted:
Issue Date:
DOI: https://doi.org/10.1007/BF02242739