21 D. Davcec and A. Burkhard. Consistency and Recovery Control for Replicated Files. In Proceedings of the 10th Symposium on Operating Systems Principles, pages 87--96, 1985.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....order to permit enough information flow between successive primary views to achieve coherence. For example, each primary view might have to contain at least a majority of the processes in the previous primary view. Several dynamic voting schemes have been developed to define primaries adaptively [12, 15, 21, 26, 33]. In particular, Lotem, Keidar, and Dolev [26] have described an implementation of a group membership service that yields only primary views, according to a dynamic notion of primary. An interesting feature of their work is that it points out various subtleties of implementing such a membership ....

D. Davcev and W. Buckhard, "Consistency and recovery control for replicated files", in ACM Syrup. on Operating Systems Principles, n.10, pp. 87-96, 1985.

....that they do not tolerate total failures. Our approach selects a primary partition based on a quorum definition that can change dynamically. Existing protocols for dynamic application driven quorum adjustments are generally tailored to specific problems, including replicated file management [DB85,JM90] transaction processing [H86,H87] emulation of multi writer multi reader registers [LS97] mutual exclusion [BGMS89, BB98] In contrast, we attempt to provide a general programming methodology in a higher level framework, i.e. membership services with (enriched) virtual synchrony. The ....

....processing [H86,H87] emulation of multi writer multi reader registers [LS97] mutual exclusion [BGMS89, BB98] In contrast, we attempt to provide a general programming methodology in a higher level framework, i.e. membership services with (enriched) virtual synchrony. The protocol in [DB85] assumes accurate and instantaneous detection of failures, as pointed out by [BGMS89] The protocol in [LS97] may be blocking, i.e. unfortunate failure of a specific node may let the protocol wait for this node to recover. The protocol in [BB98] does not tolerate partitions and assumes the ....

D. Davcev, W.A. Burkhard, "Consistency and recovery control for replicated files", Proc. 10-th ACM Symposium on Operating Systems Principles, December 1985, pp. 87-96.

....These protocols rely on quorums to provide mutual exclusion and prevent conflicting updates. As a result, they provide lower data availabilities than the other protocols. The best known quorum oriented protocols include majority consensus voting (MCV) weighted voting (WV) 5] dynamic voting (DV) [4], dynamic linear voting (DLV) 9] and voting with witnesses (VWW) 15] A common feature of all replication control protocols is the use of metadata to record the states of the replicas. These metadata nearly always include a version number, that is an integer that is incremented each time the ....

....Ascertaining the state of a replicated object requires collecting the votes of a quorum of the replicas. Should this be prevented by a sufficient number of site failures, the replicated object is considered to be inaccessible. Protocols that adjusts quorums, such as dynamic voting and its variants [4, 9], or modify the number of votes assigned to each replica [1] are known to provide higher data availability than protocol using static quorums. Cohort sets represent the set of replicas that participated in the last write operation. Whenever writes are significantly more frequent than site ....

D. Davcev and W. A. Burkhard, "Consistency and Recovery Control for Replicated Files," Proc. 10th ACM Symposium on Operating System Principles, (1985) pp. 87--96.

....on core copy management only (Section 3) In a mobile environment it is important to have the ability to reconfigure the set of replicas, for example, migrating one copy from one site to another, or adding a new copy to a set of copies. A number of such algorithms have been proposed, e.g. ET86, DB85, Her87, JM87a, JM87b, JM88, P 86, P 84] The key to understanding these algorithms is to explicitly represent the directory that specifies the sites holding copies. Then, a reconfiguration is simply a transaction that modifies the directory. Reconfiguration transaction must be processed using ....

D. Davcev and W. Burkhard. Consistency and Recovery Control for Replicated Files. In Proceedings of the Tenth ACM Symposium on Operating System Principles, pages 87--96, December 1985.

....latency. Drawback (ii) is true of all algorithms which addresses network partitioning. Finally, drawback (iii) has been addressed by some enhancements to QC which allow dynamic reconfiguration of the quorums. As a replica control protocol, QC has received much attention from researchers [1, 3, 6, 7, 13, 15, 14, 19, 8, 18] . Its implementation, however, is not widespread. 3 Protocol Implementations Replica control protocols are perhaps among the most difficult to implement in a distributed database management system. The reasons are many fold: First, due to the nature of replication there is a dramatic increase ....

D. Davcev and W. Burkhard. Consistency and Recovery Control for Replicated Files. In Proceedings of the Tenth ACM Symposium on Operating Systems Principles, pages 87--96, December 1985.

....integral part of the Echo [11] and Harp [14] file systems. Observation 4. Dynamic voting can improve the availability of replicated data managed by quorum consensus protocols. Dynamic voting protocols adjust read and write quorums whenever they detect a change in the number of available replicas [5, 12]. Central to all dynamic voting protocols, is the notion that replicas known to be unreachable should be excluded from all quorum computations. All dynamic voting protocols maintain some record of the set of replicas that are allowed to participate in elections. Because of the distributed nature ....

Davcev, D., Burkhard, W.A.: Consistency and recovery control for replicated files. Proc. 10th ACM Symp. on Operating System Principles, (1985) pp. 87--96.

....consistently. Traditionally, a primary view was defined as one containing a majority of all possible sites, but other, dynamic, definitions are possible, based on intersection properties between successive primary views. Several methods have been developed to define primary views adaptively, e.g. [6, 13, 14, 17, 27, 35, 41, 43, 47]. Producing good specifications for view oriented group communication services is difficult, because these services can be complicated, and because different such services provide different guarantees about safety, performance, and fault tolerance. Examples of specifications for group membership ....

D. Davcev and W. Buckhard. Consistency and recovery control for replicated files. In ACM Symp. on Operating Systems Principles, volume 10, pages 87--96, 1985.

....mechanism [5] 2.2 Voting protocols Voting protocols guarantee the reliability of the replicated objects by disallowing read and write accesses that cannot collect a quorum of replicas. The best known voting protocols include majority consensus voting [15] weighted voting [4] dynamic voting [1], dynamic voting, dynamic linear voting [6] and voting with witnesses [12] Each one of these protocols relies on quorum computations to provide reliability for the objects. To make such guarantees, the protocols follow two quorum requirements: W j W k j, k (1) R i W j i, j (2) where W j ....

D. Davcev and W. A. Burkhard, "Consistency and Recovery Control for Replicated Files," in Proc. 10 th ACM SOSP Symposium, 1985, pp. 87-96.

....replication control protocol, majority consensus voting (MCV) requires a minimum of three replicas to be of any practical use. Even then, quorum requirements disallow a large fraction of read and write operations. Several solutions have been proposed to surmount these limitations. Dynamic voting [DaBu85] and dynamic linear voting [JaMu87] adjust quorums to reflect changes in replica availability and network topology. Both greatly improve availability when applied to replicated objects composed of more than three replicas, but do not perform significantly better 2 than majority consensus ....

D. Davcev and W. A. Burkhard, "Consistency and Recovery Control for Replicated Files," Proc. 10th ACM Symposium on Operating System Principles (1985) pp. 87-96.

....noted that an explicit majority of votes is not required for mutual exclusion: the actual requirement is that the read quorum and update quorum each be large enough to intersect. Availability limitations imposed by infrequently accessible replicas have been addressed by dynamic voting protocols [BGS86, Her86, ES83, DB85, JM87]. These allow adjustment of quorum definitions within a majority partition, to redefine the size of a majority by effectively disenfranchising currently inaccessible replicas. Ghost replicas have been proposed as a way to reduce the actual storage costs of voting mechanisms [Par86, RT88] ....

Danco Davcev and Walter A. Burkhard. "Consistency and Recovery Control for Replicated Files." In Proceedings of the Tenth Symposium on Operating Systems Principles, pp. 87--96. ACM, December 1985.

....to be inaccessible. Majority Consensus Voting (MCV) and Weighted Voting (WV) 5] are both called static protocols because the required quorums of replicas and the number of votes assigned to each replica are fixed. Dynamic protocols that adjusts quorums, such as dynamic voting and its variants [4, 8, 10], or modify the number of votes assigned to each replica [2] can minimize the impact of site failures and increase availability. The first author is Associate Professor of Computer Science at the University of California, Santa Cruz. The Dynamic voting (DV) protocol [4] instantly adjusts ....

....and its variants [4, 8, 10] or modify the number of votes assigned to each replica [2] can minimize the impact of site failures and increase availability. The first author is Associate Professor of Computer Science at the University of California, Santa Cruz. The Dynamic voting (DV) protocol [4] instantly adjusts quorums to reflect changes in the state of the network of sites holding the replicas. The DV protocol requires each site to maintain in real time a connection vector recording the state of the network. Since the dynamic voting protocol does not assign weights to replicas and ....

[Article contains additional citation context not shown here]

D. Davcev and W. A. Burkhard, "Consistency and Recovery Control for Replicated Files," Proc. 10th ACM SOSP, (1985) pp. 87--96.

....updates are allowed only in the partition containing a majority of the replication sites. When the primary site is down, a new primary site is elected. Other recent experimental efforts in this area include RNFS at Cornell University [41] and Gemini at the University of California at San Diego [14, 52, 53]. 5.2. Scalability Certain problems induced by scale have been exposed by the extensive use of large distributed file systems. One problem is the need for decentralization. The ability to delegate administrative responsibility along lines that parallel institutional boundaries is critical for ....

Davcev, D., Burkhard, W.A. Consistency and Recovery Control for Replicated Files. In Proceedings of the 10th ACM Symposium on Operating System Principles, Orcas Island. December, 1985.

.... Recent research on management of replicated data has focused on organizing multiple copies of an object into a logical structure [2] and exploiting it to obtain smaller quorum sizes than is possible with conventional methods such as weighted voting [7] majority voting [11] dynamic voting [5] [8] and coteries [6] Examples of techniques based on logical structure are: tree protocol [1] grid protocol [4] and hierarchical quorum consensus [9] The quorum sizes are O( p N) in Maekawa s p N algorithm [10] which arranges N copies of an object in projective planes, and the grid ....

Davcev, D., and Burkhard, W., "Consistency and Recovery Control for Replicated Files," Proc. 10th ACM Symposium on Operating Systems Principles, (1985) pp 8796.

....Theory (ICDT 90) Paris, France, Dec. 1990. Lecture Notes in Computer Science 470. Berlin, Heidelberg, New York: Springer Verlag, pages 318 332 y e mail: borghoff lan.informatik.tu muenchen.dbp.de Voting Strategy Transparent access to replicated files has been extensively investigated [3, 4, 8, 10, 12, 13, 16, 17, 19, 20, 28, 29, 31, 35, 36, 39, 40, 41, 46]. The replication method proposed here is a modification of a file replication method due to Thomas [45] and Gifford [24] Our method encompasses replication methods that allow for dynamic changes of the degree of replication as well as for dynamic changes of the location of replicas [21, 26, 27, ....

D. Davcev and W.A. Burkhard. Consistency and recovery control for replicated files. In Proc. 10th ACM Symp. on Operating Systems Principles, pages 87--96, Orcas Island, WA, December 1985. ACM SIGOPS Operating Systems Review 19:5.

....order to permit enough information flow between successive primary views to achieve coherence. For example, each primary view might have to contain at least a majority of the processes in the previous primary view. Several dynamic voting schemes have been developed to define primaries adaptively [7, 10, 16, 18, 23]. In particular, Lotem, Keidar, and Dolev [18] have described an implementation of a group membership service that yields only primary views, according to a dynamic notion of primary. An interesting feature of their work is that it points out various subtleties of implementing such a mem ....

D. Davcev and W. Buckhard, "Consistency and recovery control for replicated files", in ACM Symp. on Operating Systems Principles, n.10, pp. 87--96, 1985.

....processing latency. Drawback (ii) is true of all algorithms which addresses network partitioning. Finally, drawback (iii) has been addressed by some enhancements to QC which allow dynamic reconfiguration of the quorums. As a replica control protocol, QC has received much attention from researchers [1, 9, 16, 18, 19, 29, 28, 27, 36, 20, 35] . Its implementation, however, is not widespread. The Majority Voting protocol is perhaps the version of the protocol that is typically considered. 4 Simulation Model We made use of a simulation testbed to investigate the performance of the PC and QC protocols in a client server environment. The ....

....to maintain a list of the known status of other sites in terms of their accessibility. Such a list, which we term an estimated uplist, may be maintained by having the sites periodically exchanging status verification messages, as practiced by hosts on the Internet and on the Tandem non stop system [9, 8]. In our model, we assume that such a list is available to the transaction coordinator with no additional overhead. If any site senses a failure among the server sites, a recovery process is initiated whereby data objects whose primary sites have failed may be reassigned new primary sites. If a ....

D. Davcev and W. Burkhard. Consistency and Recovery Control for Replicated Files. In Proceedings of the Tenth ACM Symposium on Operating Systems Principles, pages 87--96, December 1985.

....when there are different knowledge levels at different sites. When partitions occur, such knowledge differences are inevitable: Once a site detaches, it is impossible for other sites to know whether it received a specific message before its detachment, or not. Some past protocols (e.g. [9, 20, 11]) lead to inconsistent results in such cases, as demonstrated by the following typical scenario: ffl The systems consists of five processes: a; b; c; d and e. The system partitions into two components: a; b; c and d; e. ffl a; b and c try to form a new quorum. To this end, they exchange ....

....deciding whether a set of processes is the primary component. Like other dynamic voting protocols, our paper focuses on solving the latter problem, assuming a separate mechanism that solves the former. Our dynamic voting protocol assumes a membership mechanism no stronger than those assumed in [9, 14, 20, 16, 11, 1]. Each process is equipped with an underlying membership module, e.g. 2, 3, 12, 8] When the membership module senses failures or recoveries, it reports to the process of the new membership, i.e. the set of processes that are currently assumed to be connected. Furthermore, our protocol assumes ....

Davcev, D., and Burkhard, W. Consistency and Recovery Control for Replicated Files. In ACM Symp. on Operating Systems Principles (1985), no. 10, pp. 87-- 96.

No context found.

21 D. Davcec and A. Burkhard. Consistency and Recovery Control for Replicated Files. In Proceedings of the 10th Symposium on Operating Systems Principles, pages 87--96, 1985.

No context found.

D. Davcev and W. Burkhard. Consistency and Recovery Control for Replicated Files. In ACM Symp. on Operating Systems Principles, number 10, pages 87--96, 1985.

No context found.

Danco Davcev and Walter A. Burkhard. Consistency and Recovery Control for Replicated Files. In Proceedings of the 10th Symposium on Operating Systems Principles, pages 87-96, 1985.

No context found.

-330. b97'$d D. Davcev, W.A. Burkhard, "Consistency and recovery control for replicated files", Q...'p# ###u

No context found.

D. Davcec and A. Burkhard. Consistency and Recovery Control for Replicated Files. In Proceedings of the 10th Symposium on Operating Systems Principles, pages 87#96, 1985.

No context found.

DAVCEV,D.AND BURKHARD, W. 1985. Consistency and recovery control for replicated files. In Proceedings of the Tenth Symposium on Operating Systems Principles (Orcas Island, WA, Dec.), 87--96.

No context found.

DaBu85 D. Davcev and W. A. Burkhard, "Consistency and Recovery Control for Replicated Files," 6 Proc. 10th ACM Symposium on Operating System Principles (1985), pp. 87-96.

No context found.

DaBu85 D. Davcev and W. A. Burkhard, "Consistency and Recovery Control for Replicated Files," Proc. 10th ACM Symposium on Operating System Principles (1985), pp. 87-96.

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC