P.A. Bernstein, M. Hsu, and B. Mann, "Implementing Recoverable Requests Using Queues," in Proc. ACM SIGMOD'90, Atlantic City, pp. 112--122.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....natural style of programming is reflected in standard SQL with its session state. The problem here is losing the state when the system crashes. This may create a semantic mess that frequently requires human intervention to repair. Outages can be very long because of this. Classical TP monitors [BHM90, GrRe93] insist that applications be stateless, a rather unnatural programming style. Programming stateless workflows then requires a multi transaction application in which each step commits and passes its state in a transactional queue to the next step of the workflow. Phoenix Project The Phoenix ....

Bernstein, P.A., Hsu, M., Mann, B.: Implementing Recoverable Requests Using Queues. SIGMOD Conf., Atlantic City, NJ (1990).

....the transaction, as if the transaction executed somewhat later. After the transaction commits, the application returns to the initial state. ffl Keep application state small and write it in its entirety to a recoverable resource manager when required, e.g. database or recoverable queue [3]. This is facilitated by the use of a scripting language for the application [4, 18] The script language interpreter stores the entire application state at well chosen times so that application execution can continue from the saved state after a crash. ffl Use a persistent programming language ....

Bernstein, P., Hsu, M., and Mann, B. Implementing Recoverable Requests Using Queues. Proc. ACM SIGMOD Conf.(May, 1990) Atlantic City, NJ 112-122

....the temporal data lost during a crash event or system failure. Depending on the application, such a gap in a time series can be filled using different methods such as interpolation, data recall (if the data is stored) etc. Another alternative would be the queued transactions like technique [Bern90]. According to this approach, the remote data source locally collects the updates in a queue, and the system dequeues updates directly from there. This way, even if the system crashes, the updates will not be lost because they are being collected at a remote location. 65 In a more advanced ....

P. Bernstein, M. Hsu, and B. Mann, "Implementing recoverable requests using queues", In Proceedings of the ACM-SIGMOD International Conference on Management of Data, pp. 112-122, Atlantic City, N.J., May 1990.

....constructors will also make it easier to implement the persistent programming languages discussed in Proposition 3.2. Furthermore, as time unfolds it is certainly possible that additional type constructors may become desirable. For example, transaction processing systems manage queues of messages [BERN90]. Hence, it may be desirable to have a type constructor which forms queues. Second generation systems have few of these type constructors, and the advocates of Object oriented Data Bases (OODB) claim that entirely new DBMSs must come into existence to support these features. In this regard, we ....

Bernstein, P. et. al., "Implementing Recoverable Requests Using Queues", Proc. ACM SIGMOD Conference on Management of Data, Atlantic City, N.J., May 1990.

....For removal of forwarding addresses, or for reclustering, one may wish to keep a request queue. One does not wish the queue to disappear if there is a system failure. So this should be a recoverable queue. Many transaction processing systems implement recoverable queues [Bernstein1990a] [Bernstein1990b]. The record moving transaction would begin by removing a request from the queue. If a record moving transaction aborts, or if a system failure occurs before the record moving transaction can complete, the UNDO of this step would place the request back in the queue. Since requests can be ....

P. Bernstein, M. Hsu, B. Mann. Implementing Recoverable Requests Using Queues. Proc. ACM SIGMOD Conference on Management of Data. 1990 pp.112-122.

....semantics of the rules that are triggered during the execution of a workflow (or an ATM) are also defined within the ATM model. The model supports extended recovery options of roll back and compensation as well as forward recovery. An implementation of the ATM model based on persistent queues [AMB90] under the assumption that the system supports nested transactions is also proposed. While the ATM model supports dynamic workflows, similar to the extended transaction models, its implementation requires coordinated execution among the processing stations. Furthermore, the implementation ....

Bernstein P. A., Hsu M., and Mann B. Implementing recoverable requests using queues. In Proceedings of ACM-SIGMOD 1990 International Conference on Management of Data, Atlantic City, New Jersey, 1990.

....uncommitted event signal suffers from the dirty dependent operation 1 (DDO) problem. The extended version of Ariel being developed avoids this problem by making event transmission depend on commit of the transaction raising the event. As part of the solution to these problems, recoverable queues [BHM90] are used to help communicate events from servers back to clients. Recoverable queues and other techniques have been used in commercial products for many years to ensure recoverable message delivery in distributed systems [GR93] The goal of this work is to provide an integrated framework for ....

....Issues If a client application fails and then comes back up, then when the application restarts, it must reconnect to its durable event queue if it is expecting to receive any event signals via this queue. Reconnecting is done 9 via a variation of the normal mechanism used for durable queues [BHM90]. If the database system fails and comes back up, then the normal recovery mechanism will restore all the durable queues to a transactionconsistent state. Once the durable queues have been recovered, applications can start taking events out of them. If the network connecting the client and the ....

[Article contains additional citation context not shown here]

Philip A. Bernstein, Meichun Hsu, and Bruce Mann. Implementing recoverable requests using queues. In Proceedings of the ACM SIGMOD International Conference on Managment of Data, pages 112--122, May 1990.

....are propagated to the replicas in MSets. Each local system is responsible for applying its MSet and preserving internal consistency. Note that the propagation of MSets to each site is asynchronous. We assume the system maintains the unprocessed MSets in some stable storage, such as stable queues [5] and persistent pipes [16] Each MSet is stored as an element in a stable queue. Due to the asynchronous propagation of MSets, replicas of a logical object can differ at any given moment. This is the source of inconsistency seen by the query ETs. A key observation, however, is that under ESR all ....

P.A. Bernstein, M. Hsu, and B. Mann. Implementing recoverable requests using queues. In Proceedings of 1990 SIGMOD International Conference on Management of Data, pages 112--122, May 1990.

....of reliably storing some message recovery information . In the context of exactly once transaction processing, this recovery information should represent the transaction processing state. Several approaches were proposed in the literature to store that state for recovery purposes, e.g. [7, 8, 9]. Nevertheless, those approaches do not guarantee the high availability of that state. Furthermore, they rely on disk storage at the client or at some application server. Relying on the client s disk is problematic if the client is a Java applet that does not have the right to access the disk. ....

P. Bernstein, M. Hsu, and B. Mann, "Implementing recoverable requests using queues," in Proceedings of the 1990 ACM SIGMOD International Conference on Management of Data, May 1990.

....that was triggered will, in fact, complete in time to assure the consistency of the data. To allow the parent transaction some degree of control over the execution of a child transaction, the concept of a VMS mailbox has been generalized in [GMGK 90] Similar ideas have been presented in [BHM90] and in [HS90] where the notion of a persistent pipe has been introduced. Both generalized mailboxes and persistent pipes allow the parent transaction to send a message to a child process and know that the message will be eventually delivered. If such a guarantee is sufficient, the parent ....

P. Bernstein, M. Hsu, and B. Mann. Implementing Recoverable Requests Using Queues. In Proceedings of ACM SIGMOD Conference on Management of Data, 1990.

....between p 1 and p 2 , as long as we can guarantee the p 2 will eventually commit once p 1 commits. To guarantee that the subsequent pieces of a distributed transaction will eventually commit once the first piece commits, we adopt the mechanism of persistent transmission [42] or recoverable queues [16], which are offered in commercial transaction mon 63 itors. A recoverable queue is an inter site communication channel, through which data are guaranteed to survive site failures as well as link failures. Unlike other types of inter site communication channels, both the sender and the receiver ....

P. Bernstein, M. Hsu, and B. Mann. Implementing recoverable requests using queues. In Proceedings of 1990 SIGMOD International Conference on Management of Data, pages 112--122, May 1990.

....any existing data model, like sessions or windows. Context Management In principle, there are three different ways to manage context reliably: a) keeping it in the global database; b) transferring it explicitly from one step to another, e.g. through a reliable queue mechanism [GaMo91, GGKKS91, BeHM90] c) setting up a special context database with a private interface for each ConTract. The first possibility would require the step programmer to know about the public context database, its structure and how to access the needed context elements. Apart from complicating the step code ....

Bernstein, P.A., Hsu, M., and Mann, B. Implementing Recoverable Requests Using Queues. Proc. ACM SIGMOD, 1990.

No context found.

P.A. Bernstein, M. Hsu, and B. Mann, "Implementing Recoverable Requests Using Queues," in Proc. ACM SIGMOD'90, Atlantic City, pp. 112--122.

No context found.

P. Bernstein, M. Hsu, and B. Mann. Implementing recoverable requests using queues. In SIGMOD Conference, pages 112--122, 1990.

No context found.

Philip A. Bernstein, Meichun Hsu, and Bruce Mann. Implementing recoverable requests using queues. In Hector GarciaMolina and H. V. Jagadish, editors, Proceedings of the 1990 ACM SIGMOD International Conference on Management of Data, Atlantic City, NJ, May 23-25, 1990, pages 112--122. ACM Press, 1990.

No context found.

Philip A. Bernstein, Meichun Hsu, and Bruce Mann. Implementing recoverable requests using queues. In Hector Garcia-Molina and H. V. Jagadish, editors, pages 112--122. ACM Press, 1990.

No context found.

P. Bernstein, M. Hsu and B. Mann, \Implementing Recoverable Requests Using Queues", Proc. ACM Int. Conference on Management of Data (SIGMOD), 1990.

No context found.

P. Bernstein, M. Hsu and B. Mann, "Implementing Recoverable Requests Using Queues", Proc. ACM Int. Conference on Management of Data (SIGMOD), 1990.

No context found.

P. A. Bernstein, M. Hsu, and B. Mann. Implementing recoverable requests using queues. In H. Garcia-Molina and H. V. Jagadish, editors, Proceedings of the 1990 ACM SIGMOD International Conference on Management of Data, Atlantic City, NJ, May 23-25, 1990, pages 112--122. ACM Press, 1990.

No context found.

-- Bernstein, P., Hsu, M. and Mann, B, "Implementing Recoverable Requests Using Queues.' 1990 ACM SIGMOD Conference, Atlantic City, May 1990

No context found.

P. A. Bernstein, M. Hsu, and B. Mann. Implementing recoverable requests using queues. In ACM SIGMOD, pages 112--122, 1990.

No context found.

P. A. Bernstein, M. Hsu, and B. Mann. Implementing Recoverable Requests Using Queues. In SIGMOD, 1990.

No context found.

Philip A. Bernstein, Meichun Hsu, and Bruce Mann. Implementing recoverable requests using queues. ACM SIGMOD, pages 112--122, 1990.

No context found.

-- Bernstein, P., Hsu, M. and Mann, B, "Implementing Recoverable Requests Using Queues.' 1990 ACM SIGMOD Conference, Atlantic City, May 1990

No context found.

P.A. Bernstein, M. Hsu and B. Mann. "Implementing Recoverable Requests Using Queues." Proc.. ACM SIGMOD. Atlantic City, NJ. 1990

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