D. Long, The Management of Replication in a Distributed System, num. UCSL-CRL-88-07, University of California, Santa Cruz, CA, 1988.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....the major challenge, to tolerate network partitions. One of the most important feature of the Goofy service is to provide higher availability while ensuring object consistency. These two goals can be achieved in a partition free network (as with the Optimistic Available Copy Protocol described in [8]) but it is impossible to guarantee the object consistency in presence of network partitions without decreasing the availability. Some trade offs have to be found to satisfy Goofy s requirements. 3.2.2 Consistency Requirements To guarantee that jobs share the same image of an object, we could ....

....ensure that Goofy is able to guarantee the strict consistency of objects, we use a pessimistic policy. Such a policy can be achieved using a quorum consensus algorithm. The quorum is a simple pessimistic method which allows to resist to network partitions during the collection of the votes (see [5][8] for more details) The following example illustrates this protocol: Suppose that the GMS1 wants to load the object 01. GMS1 sends a Load request to the replica R1. So, R1 is called the initiator of the lock. It sends a Prepare to lock request to the others replicas (i.e. R2 and R3) and then it ....

[Article contains additional citation context not shown here]

D. Long, The Management of Replication in a Distributed System, num. UCSL-CRL-88-07, University of California, Santa Cruz, CA, 1988.

....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.D.E. Long. The management of replication in a distributed system. Technical Report UCSC/CRL 88/07, Dept. of Computer and Information Science, Univ. of California, Santa Cruz, CA, 1988.

....can be used to implement arbitrary replication and consistency protocols on top of transparent distributed shared memory [LH89] The contribution of FI is the ability to execute application code before and after memory pages are accessed. This code might, for example, implement a voting algorithm [Lon88] The consistency protocol runs transparently; the client accesses the memory with normal load and store instructions. 3.3. Consistency and Replication Control 9 One possible implementation is the following. Shared memory pages are replicated on all the participating sites. Upon a write, the ....

Darrell D. E. Long. The Management of Replication in a Distributed System. Ph.D. dissertation, University of California at San Diego, August 1988.

....is similar except that the primary copy of a particular data object will be chosen independently of other data objects. With available copy, writes are sent to every copy of a data object, and reads are performed on any copy available. Failures are detected by sites that remain in operation [28]. In all of these protocols, there are significant weaknesses that adversely affect the robustness of a database. Both the primary site and primary copy protocols depend upon a single copy for all data updates on particular data objects. The only advantage is that reads can be made from any site. ....

....witnesses. Witnesses are not data objects; they are records of the state of the data object [33] A witness must be a part of the quorum when the voting entity reads data objects. Other modifications of these simple algorithms include dynamic vote adjustment [4] and dynamic quorum adjustment [28] protocols. These algorithms make a distributed database more sensitive to changes in its configuration, and this increases the robustness of the database. Improving the voting algorithm in the distributed database is another goal for our package in the future. 2.2.6 Transport level data ....

[Article contains additional citation context not shown here]

D. D. E. Long. The Management of Replication in a Distributed System. PhD thesis, University of California at San Diego, 1988. 66

....that is C (W s ) n [ k=0 C k (W s ) where C k (W s ) S t2Ws C k Gamma1 (W t ) and C 0 (W s ) W s . 3. A More Efficient Available Copy Protocol One of the major objectives of the OAC protocol was to reduce the costs of updating the was available sets of operational sites [11]. So it was decided that: 1. was available set updates should always be piggy backed on existing read, write and site recovery operations, and 2. was available set updates should never involve sites that were not involved in each read, write or site recovery operation. Hence, site recovery ....

....have had the two advantages of (a) making all available sets current and (b) removing the need to compute the closure of these sets every time the system has to recover from a total failure. As it happened, the OAC protocol was formalized [10] well before its data availability was fully analyzed [11, 16]. So the benefits of updating the was available sets at recovery time were only understood after the protocol had been fully specified and this update was done independently of the site recovery process itself [11, 16] An even more important simplification could be achieved if the was available ....

[Article contains additional citation context not shown here]

D. D. E. Long, "The Management of Replication in a Distributed System," Ph.D. dissertation, University of California, San Diego, 1988.

....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 ....

....apply although the read quorum mentioned in Lemma 1 and in Theorem 1 must now be replaced by a majority of the replicas belonging to the current majority block. The simplicity of having to maintain only one cohort set can be contrasted with the complexity of the optimistic dynamic voting protocol [10] where each replica must maintain: 1. a partition set P i representing the set of sites which participated in the last successful operation on the replicated data, 2. an operation number, o i that is incremented at every access, and 3. a version number, v i that is incremented at every write. ....

[Article contains additional citation context not shown here]

D. D. E. Long, "The Management of Replication in a Distributed System," Ph.D. dissertation, University of California, San Diego, 1988.

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