A. Adya. Weak Consistency: A Generalized Theory and Optimistic Implementations for Distributed Transactions. PhD thesis, Massachusetts Institute of Technology, Cambridge, MA, Mar. 1999. Also available as Technical Report MIT/LCS/TR-786.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....high so that it becomes impossible for using the system. The strong consistency algorithm [5, 8, 9] is suitable for systems with few replicas, and on a reliable network where a large amount of bandwidth is available. However, weak consistency algorithms, i.e. Golding s Refdbms [12] Thesis s Adya [4], Usenet [14, 22] Coda [23] Bayou [21] Ficus [13] Grapevine [24] generate very little traffic, low latency, and are more scalable. They do not sacrifice either availability or reply time in order to guarantee strong consistency, but only need to ensure that the replicas eventually converge to ....

Adya, "Weak Consistency: A Generalized Theory and Optimistic Implementations for Distributed Transactions", PhD thesis MIT, Department of Electrical Engineering and Computer Science, March 1999.

....there never need to be more than three versions of the same block, and puts bounds on the costs of cleaning, or garbage collecting, versions that are no longer needed. 7.5. 3 Optimistic Protocols In the same vein as transient versioning, there are a number of schemes in optimistic concurrency [Amiri99, Adya99, Kung81] that would apply to an Active Disk setting. Such protocols maintain a list of blocks that a particular transaction depends on, and a list of blocks to be modified. The system then uses write ahead logging to ensure that all drive operations are recoverable in the case that they are aborted at ....

Adya, A. "Weak Consistency: A Generalized Theory and Optimistic Implementations for Distributed Transactions" PhD Thesis, MIT Laboratory for Computer Science, March 1999.

....[11] and Oracle s Snapshot Isolation and Read Consistency [24] and new levels; for example, we have developed an additional isolation level called PL 2 , which is the weakest level that guarantees consistent reads and causal consistency with respect to transactions. Details can be found in [1]. Our definitions are given using a combination of constraints on transaction histories and graphs; we proscribe different types of cycles in a serialization graph at each isolation level. Our graphs are similar to those that have been used before for specifying serializability [9, 19, 14] ....

....the history only shows reads of versions that were actually observed by T i . 4.3.2 Predicate based Modifications A modification based on a predicate P is modeled as a predicate based read followed by write operations on tuples that match P. Although this approach is weaker than the one used in [1], it models the behavior of commercial databases. For example, suppose transaction T i executes the following code for the employee database discussed above: UPDATE EMPLOYEE SAL = SAL 10 WHERE DEPT=SALES; Suppose that the system selects versions, x 1 , y 2 , and z init for this operation. If ....

[Article contains additional citation context not shown here]

A. Adya. Weak Consistency: A Generalized Theory and Optimistic Implementations for Distributed Transactions. PhD thesis, MIT, Cambridge, MA, Mar. 1999.

....the delay to three messages for read write transactions. For read only transactions, participants send their decision directly to the client, so that there are just two message delays. Furthermore, we have developed a way to commit read only transactions entirely at the client almost all the time [Ady99] Now we discuss how validation works. We use backward validation[Hae84] the committing transaction is compared with other committed and committing transactions but not with active transactions (since that would require additional communication between clients and servers) A transaction s ....

A. Adya. Weak Consistency: A Generalized Theory and Optimistic Implementations for Distributed Transactions. PhD thesis, Massachusetts Institute of Technology, Cambridge, MA, Mar. 1999.

No context found.

A. Adya. Weak Consistency: A Generalized Theory and Optimistic Implementations for Distributed Transactions. PhD thesis, Massachusetts Institute of Technology, Cambridge, MA, Mar. 1999. Also available as Technical Report MIT/LCS/TR-786.

No context found.

A. Adya. Weak Consistency: A Generalized Theory and Optimistic Implementations for Distributed Transactions. PhD thesis, Massachusetts Institute of Technology, Cambridge, MA, Mar. 1999. Also available as Technical Report MIT/LCS/TR-786.

No context found.

A. Adya. Weak Consistency: A Generalized Theory and Optimistic Implementations for Distributed Transactions. PhD thesis, Massachusetts Institute of Technology, Cambridge, MA, March 1999.

No context found.

A. Adya. Weak Consistency: A Generalized Theory and Optimistic Implementations for Distributed Transactions. PhD thesis, Massachusetts Institute of Technology, Cambridge, MA, March 1999.

No context found.

A. Adya, "Weak Consistency: A Generalized Theory and Optimistic Implementations for Distributed Transactions", PhD thesis Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, March 1999.

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