Divyakant Agrawal
Amr El Abbadi
Abstract
In this paper, we present an efficient and fault-tolerant algorithm for generating quorums to solve the distributed mutual exclusion problem. The algorithm uses a logical tree organization of the network to generate tree quorums, which are logarithmic in the size of the network in the best case. Our approach is resilient to both site and communication failures, even when such failures lead to network partitioning. Furthermore, the algorithm exhibits a property of graceful degradation, i.e., it requires more messages only as the number of failures increase in the network. We describe how tree quorums can be used for various distributed applications for providing mutually exclusive access to a distributed resource, managing replicated objects, and atomically commiting a distributed transaction.
- 1 AGRAWAL, D., AND EL ABBADI, A. Exploiting logical structures of replicated databases. Inf. Process. Lett. 33, 5 (Jan. 1990), 255-260 Google Scholar
- 2 AGRAWAL, D., AND EL ABBADI, A. The tree quorum protocol: An efficient approach for managing replicated data. In Proceedings of Sixteenth International Conference on Very Large Data Bases (Aug. 1990), 243-254. Google Scholar
- 3 ALSBERG, P. A., AND DAY, J.D. A principle for resilient sharing of distributed resources. In Proceedings of the Second International Conference on Software Engineering (Oct. 1976), 562-570. Google Scholar
- 4 BERNABI~u-AUBAN, J. M., AND AHAMAD, M. Applying a path-compression technique to obtain an efficient distributed mutual exclusion algorithm. In Proceedings of the Third International Workshop on Distributed Algorithms (Sept. 1989), 33-44. Google Scholar
- 5 DAVIDSON, S. B., GARCIA-MOLINA, H., AND SKEEN, D. Consistency in partitioned networks. ACM Comput. Surv. 17, 3 (Sept. 1985), 341-370. Google Scholar
- 6 GARCIA-MOLINA, H., AND BARBARA, D. How to assign votes in a distributed system. J. ACM 32, 4 (Oct. 1985), 841-860. Google Scholar
- 7 GIFFORD, D.K. Weighted voting for replicated data. In Proceedings of the Seventh A CM Symposium on Operatzng Systems Prznciples (Dec. 1979), 150-159. Google Scholar
- 8 GRAY, J. N Notes on database systems. In Operating Systems: An Advanced Course, vol. 60 of Lecture Notes in Computer Science. R. Bayer, R. M. Graham, and G. Seegmuller, Eds. Springer-Verlag, New York, 1978, 393-481 Google Scholar
- 9 HELARY, J. M., PLOUZEAU, N., AND RAYNAL, M. A distributed algorithm for mutual exclusion in an arbitrary network. Computer J. 31, 4 (1988), 289-295. Google Scholar
- 10 LAMPORT, L. Time, clocks, and the ordering of events in a distributed system. Commun. ACM 21, 7 (July 1978), 558-565. Google Scholar
- 11 LYNCH, N A., AND MERRITT, M . Introduction to the theory of' nested transactions. Tech Rep MIT-LCS-TR-367, MIT, Cambridge, Mass., 1986. Google Scholar
- 12 MAEKAWA, M . A v n algorithm for mutual exclusion in decentralized systems. A CM Trans. Comput. Syst. 3, 2 (May 1985), 145-159. Google Scholar
- 13 MAHANEY, S R., AND SCHNEIDER, F. B. Inexact agreement: Accuracy, precision, and graceful degradation. In Proceedings of the Fourth ACM Symposium on Principles of Distmbuted Computing (Aug. 1985), 237-249. Google Scholar
- 14 MJSRA. J. Detecting termination of distributed computations using markers. In Proceedings of the Second ACM Symposium on Princzples of Distributed Computing (Aug. 1983), 290-294. Google Scholar
- 15 Moss, J E B. Nested Transactzons: An Approach to Rehable Dzstrzbuted Computtng MIT Press, Cambridge, Mass., 1985. Google Scholar
- 16 NAIMI, M., AND TREHAL, M. How to detect a failure and regenerate the token m the log n distributed algorithm for mutual exclusion. In Proceedings of the Second Internatmnal Workshop on Dtstributed Algorithms, Lecture Notes in Computer Scteuce 312, Springer- Verlag, New York, 1987, 155-166 Google Scholar
- 17 RAYMOND, K A tree-based algorithm for distributed mutual exclusion. ACM Trans Comput. Syst. 7, 1 (Feb. 1989), 61-77. Google Scholar
- 18 RICARW, G., AND AC, RAWALA, A K An optimal algorithm for mutual exclumon m computer networks Commun. ACM 24, 1 (Jan. 1981), 9-17. Google Scholar
- 19 SANDEaS, B. A The information structure of dlstributed mutual exclusion algorithms ACM Trans. Comput. Syst. 5 (Aug. 1987), 284-299 Google Scholar
- 20 SCHNEIDER, F B Synchronization in distributed programs. A CM Trans. Program Languages Syst. 4, 2 (April 1982), 125-148 Google Scholar
- 21 SK~EN, D. Non-blocking commit protocols. In Proceedings of the A CM SIGMOD Conference on Management of Data (June 1982), 133-147 Google Scholar
- 22 SKEIN, D A quorum based commit protocol. In Procee&ngs of the 6th Berkeley' Workshop on D~strtbuted Data Management and Computer Networks (Feb. 1982), 69-80Google Scholar
- 23 SuzuKi, I, A~D KaSAMI, T. A distributed mutual exclusion algorithm A CM Trans Cornput. Syst. 3, 4 (Nov 1985), 344-349 Google Scholar
- 24 TANt,:NBAU~I, A S. Computer Networks 2d ed. Prentice Hall, Englewood Cliffs, N J, 1988. Google Scholar
- 25 THOMAS, R H. A majority consensus approach to concurrency control for multiple copy databases. ACM Trans. Databa,~e Syst. 4, 2 (June 1979), 180-209. Google Scholar
- 26 VAN DE S~Et'SCHEUT, J. L Fair mutual exclusmn on a graph of processes Dzstrtbut Comput 2 (1987), 113-115.Google Scholar
Index Terms
- An efficient and fault-tolerant solution for distributed mutual exclusion
Recommendations
A Token-Based Fault-Tolerant Distributed Mutual Exclusion Algorithm
In this paper, we propose a token-based distributed mutual exclusion algorithm that is resilient to site and communication failures. The protocol uses the notion of logical time to detect the loss of the token and to recover the state of the lost token. ...
A survey of permission-based distributed mutual exclusion algorithms
The problem of mutual exclusion in distributed systems has attracted considerable attention over the last two decades. The mutual exclusion problem requires that, at a time, only one of the contending processes be allowed to enter its critical section (...
A fault-tolerant h-out of-k mutual exclusion algorithm using cohorts coteries for distributed systems
PDCAT'04: Proceedings of the 5th international conference on Parallel and Distributed Computing: applications and TechnologiesIn this paper, we propose a distributed algorithm for solving the h-out of-k mutual exclusion problem with the aid of a specific k-coterie — cohorts coterie. The proposed algorithm is resilient to node and/or link failures, and has constant message cost ...
Reviews
David James Taylor
The authors have developed a novel technique for achieving mutual exclusion in distributed systems. They provide sufficient background to make the paper accessible to readers not familiar with previous work in the area. The novelty of their technique is in using a logical binary-tree structure and basing quorums on a root-to-leaf path in the tree. In this way, they can achieve mutual exclusion by contacting as few as log n nodes in an n -node system. Their quorum-construction technique can also tolerate failed nodes, building more complicated quorum sets using the binary-tree structure. Their technique achieves significantly lower costs than a majority-consensus algorithm. The costs are, of course, higher than in a centralized scheme. The tolerance of failures is difficult to quantify, and the authors provide graphs indicating that, for most situations of interest, the observed tolerance will be close to that of a majority-consensus algorithm. Their technique thus provides a valuable new tradeoff between costs and tolerance of failures. In comparing their algorithm with others, the authors make the unfortunate claim that their algorithm does not require nodes to be selected for special purposes based on their superior reliability. In fact, as they admit later in the paper, it is desirable for nodes near the root to be more reliable than nodes near the leaves. The paper does not contain any analysis of the effects of nonuniform node reliability on the algorithm. One application presented for the new technique is the handling of replicated data. In such cases, providing small quorums for read operations at the expense of large quorums for write operations is often desirable. The technique, as presented, requires th at all quorums be constructed identically, which prohibits such a tradeoff.
Access critical reviews of Computing literature here
Become a reviewer for Computing Reviews.
Comments