Muth, P., Rakow, T.C.: Atomic Commitment for Integrated Database Systems. In: Proceedings of IEEE 7th Int. Conf. on Data Engineering, Kobe, Japan (1991).  Home/Search   Document Details and Download   Summary   Related Articles   Check

....these transactions in order to achieve a high degree of concurrency. Recovery is based on inverse operations, compensating the changes of transactions which have to be aborted. Inverse operations are also used to ensure atomic commitment. Unilateral aborts as well as system crashes can be handled [25, 23]. 2.7 Tailoring the Data Model One major benefit claimed for object oriented database systems is extensibility. Although object oriented data models allow users to define their own classes, they come with a fixed set of data model primitives. We consider this a major problem as different ....

Muth, P., Rakow, T.C.: Atomic Commitment for Integrated Database Systems. In: Proceedings of IEEE 7th Int. Conf. on Data Engineering, Kobe, Japan (1991).

....these transactions in order to achieve a high degree of concurrency. Recovery is based on inverse operations, compensating the changes of transactions which have to be aborted. Inverse operations are also used to ensure atomic commitment. Unilateral aborts as well as system crashes can be handled [25, 23]. 2.7 Tailoring the Data Model One major benefit claimed for object oriented database systems is extensibility. Although object oriented data models allow users to define their own classes, they come with a fixed set of data model primitives. We consider this a major problem as different ....

Muth, P., Rakow, T.C.: Atomic Commitment for Integrated Database Systems. In: Proceedings of 1EEE 7th Int. Conf. on Data Engineering, Kobe, Japan (1991).

....multidatabase. The ticket is successively incremented by each global subtransaction, with the increment order following the dictated serialization order. Other possible schemes include [24, 23, 17] For atomic commitment, we follow the local commitment before global decision algorithm defined in [20]. Another commit scheme is defined in [16] 9 4.2 Scheduling Actions and Steps At any specific time, the different in process Interactions in the database have one or more concurrent threads executing the different tasks and subtasks. For scheduling purposes, the Interaction Manager maintains ....

....[13] were the first to really explore issues concerning multidatabase transaction serializability, and they came up with a list of requirements that need to be met by the global transaction manager. A thorough survey of the approaches to these problems can be found in [2] Specific work includes [20, 16, 24, 23, 12, 17]. Elmagarmid and Du [9] explore how different local concurrency control models affect their serialization order. Schemes that do not necessarily enforce serializable global transaction executions, but try to approximate it, include quasi serializability [7] also called multidatabase ....

Peter Muth and Thomas G. Rakow. Atomic commitment for integrated database systems. In 1991 Data Engineering Proceedings, pages 296--304, 1991. 29

....of the partially committed global transactions, in the compensate approach, attempt is made to roll back the effects of such a transaction by executing compensating transactions to undo the effects of the committed subtransactions. The compensate approach has been studied in [Levy et al. 1991; Muth and Rakow 1991; Mehrotra et al. 1992b] The compensate approach depends upon the existence of a suitable compensating transaction to undo the effects of the committed subtransaction. Notice that since the effects of the committed subtransaction may have been observed by other transactions before the ....

Muth, P. and Rakow, T. C. 1991. Atomic commitment for integrated database systems. In Proceedings of the Seventh International Conference on Data Engineering, Kobe, Japan (April 1991).

....transactions. Preserving the atomicity or semantic atomicity [GM83] of global transactions in the HDDBS systems has been recognized as an open and dif cult issue [SSU91] The traditional two phase commit protocol (2PC) developed in distributed database environments has been shown [LKS91a, SKS91, MR91, BZ94] to be inadequate to the preservation of the atomicity of global transactions in the HDDBS environment. For example, some local database systems may not support a visible prepare to commit state, in which a transaction has not yet been committed but is guaranteed the ability to commit. In ....

P. Muth and T.C. Rakow. Atomic commitment for integrated database systems. In Proceedings of the 7th International Conference on Data Engineering, pages 296-304, Kobe, Japan, April 1991.

....transactions. Preserving the atomicity or semantic atomicity [GM83] of global transactions in the HDDBS systems has been recognized as an open and difficult issue [SSU91] The traditional two phase commit protocol (2PC) developed in distributed database environments has been shown [LKS91a, SKS91, MR91, BZ94] to be inadequate to the preservation of the atomicity of global transactions in the HDDBS environment. For example, some local database systems may not support a visible prepare to commit state, in which a transaction has not yet been committed but is guaranteed the ability to commit. In ....

P. Muth and T.C. Rakow. Atomic commitment for integrated database systems. In Proceedings of the 7th International Conference on Data Engineering, pages 296--304, Kobe, Japan, April 1991.

....multidatabase. The ticket is successively incremented by each global subtransaction, with the increment order following the dictated serialization order. Other possible schemes include [20, 19, 16] For atomic commitment, we follow the local commitment before global decision algorithm defined in [18]. Another commit scheme is defined in [15] 4.2 Scheduling Actions and Steps At any specific time, the different in process Interactions in the database have one or more concurrent threads executing the different tasks and subtasks. For scheduling purposes, the Interaction Manager maintains ....

....[12] were the first to really explore issues concerning multidatabase transaction serializability, and they came up with a list of requirements that need to be met by the global transaction manager. A thorough survey of the approaches to these problems can be found in [2] Specific work includes [18, 15, 20, 19, 11, 16]. Elmagarmid and Du [8] explore how different local concurrency control models affect their serialization order. Schemes that do not necessarily enforce serializable global transaction executions, but try to approximate it, include quasi serializability [6] also called multidatabase ....

Peter Muth and Thomas G. Rakow. Atomic commitment for integrated database systems. In 1991 Data Engineering Proceedings, pages 296-- 304, 1991.

....re execute it until commitment or undo the effects of the committed subtransactions. The strategies characterizing these approaches can be classified by the relative timing of the commitment of subtransactions in the local databases with respect to the global transaction commit abort decision [MR91] The work in [WV90, BST90] enforces a global decision on the subtransactions by redoing or retrying them as necessary. The work in [PRR91, LKS91b, NZ94] commits subtransactions locally before a global decision is made and rely on compensation when a global transaction is aborted. The work in ....

P. Muth and T.C. Rakow. Atomic commitment for integrated database systems. In Proceedings of the 7th International Conference on Data Engineering, pages 296--304, Kobe, Japan, April 1991.

....through such conflicts. Thus, in general, the retry technique can be used for ensuring atomicity of transactions under the restrictions that subtransactions do not have data dependencies and that each subtransaction is retriable. The above scheme for ensuring atomicity was first mentioned in [45]. It is clear that due to restrictions that need to be imposed upon the transactions, the retry approach by itself is of limited applicability. However, as we will see later since it is possible to use each of the approaches discussed in conjunction, it may provide us with a powerful model. In any ....

....aborted, and compensating transactions are run for all the subtransactions that did commit. In the common case where subtransactions are successful, the O2PC lets sites commit sooner than in the 2PC protocol, leading to improved performance. The O2PC protocol was also developed independently in [45]. Processing distributed transactions without an atomic global commit protocol is also studied in [34] It must be noted that these commit protocols do not require each local DBMS to support a prepared state for commitment of multi site transactions and are thus attractive for MDBS environments. ....

P. Muth and T. C. Rakow. Atomic commitment for integrated database systems. Proceedings of the Seventh International Conference on Data Engineering, Kobe, Japan, April 1991.

....Both in [2] and [3] a simulated prepared state and a resubmission was proposed to deal with some failures, but the unilateral aborts in the prepared state were not included. The idea of resubmission may be found also in [14] to serve the needs of site recovery but not subtransaction recovery. In [18] various schemes are proposed, some of them guaranteeing serializability and some of them not, yet a special global scheduler (e.g. lock based) has to be used to maintain serializability. Neither local transactions are allowed in the system nor unilateral aborts are taken care of. A notable ....

P. Muth and T. Rakow, "Atomic Commitment for Integrated Database Systems", Proc. 7th Conf. Data Eng. (Kobe, 8-12 April, 1991), pp. 296-304.

....modifying globally updatable data items would not be allowed to read locally updatable items. A similar solution is used by [20] in their 2PC Agent Method method. The drawback of this kind of approaches is that they restrict access to the objects in the databases. The retry approach proposed in [12] involves re running a failed subtransaction (if another one committed at a different site) However, as pointed out in [3] one needs to make sure that any values the original subtransaction reads were not passed to other subtransactions; it must also be true that the subtransaction is repeatable ....

P. Muth and T.C. Rakow. "Atomic Commitment for Integrated Database Systems," Proceedings of the 7th Intl. Conference on Data Engineering, pp. 296-304, 1991.

....set by a EST into retained locks when the EST terminates. Retained locks are ignored by ESTs, but must still be observed by LTs. For recovery, we use the simplest variant of the multilevel recovery mechanisms developed in [Wei91, WHBM90, WH93] which has been adopted to an FDBMS environment in [MR91, MRKN92]. Recovery for LTs and for ESTs is actually provided by the DBMSs, since both LTs and ESTs are executed as regular transactions in the underlying DBMS. Note that ensuring atomicity of ESTs is an important prerequisite for the compensation based higher level recovery of GTs [WHBM90, WH93] ....

Muth, P., Rakow, T., Atomic Commitment for Integrated Database Systems, IEEE Data Engineering Conference, 1991.

....similar to log sequence numbers, are needed between the two layers of recovery to test the applicability of a log record and to ensure the idempotence of recovery steps. Finally, this recovery approach does still need a distributed commit protocol on top of the underlying database systems [WV90, MR91], since top level operations are not necessarily compatible with each other. Consequently, the semantic locks for the top level operations must be held until the end of the global commit protocol. While this may first appear as a severe impact on the execution autonomy of the participating ....

Muth, P., Rakow, T.C., Atomic Commitment for Integrated Database Systems, IEEE Data Engineering Conference, 1991

....re execute it until commitment or undo the effects of the committed subtransactions. The strategies characterizing these approaches can be classified by the relative timing of the commitment of subtransactions in the local databases with respect to the global transaction commit abort decision [11]. 18, 3] enforce a global decision on the subtransactions by redoing or retrying them as necessary. 14, 9, 13] commit subtransactions locally before a global decision is made and rely on compensation when a global transaction is aborted. 10, 12] combine these two approaches. With the forward ....

P. Muth and T. Rakow. Atomic commitment for integrated database systems. In Proceedings of the 7th International Conference on Data Engineering, pages 296--304, Kobe, Japan, Apr. 1991.

.... have been proposed for a failure free environment [BST87, BS88, Pu87, Pu88, ED90, LE90, GRS91, MRKS92, BRG92, HHS93] Recently, researchers have addressed the issue of transaction management in a failureprone environment and a number of proposal have been made [BST90, WV90, BO91, Geo91, NSS91, MR91, BS92, VW92] Each proposed recovery algorithm imposes some restrictions, which affect different aspects of local autonomy. These restrictions include: ffl limiting the types of concurrency control mechanisms used by local DBMSs. All FDBS recovery algorithms in the literature have some limitation ....

.... use by the FDBS immediately after site failure [BO91, Geo91] ffl prohibiting local DBMSs from aborting a transaction after all operations of the transaction are completed [Geo91] ffl giving up control autonomy, i.e. no local transactions can be executed without involving the FDBS [NSS91, MR91] The objective of transaction execution in centralized or distributed database systems is to achieve serializability [EGLT76] However, transaction execution in a failure prone FDBS environment is different from that in centralized or distributed database systems. To achieve atomic commitment of ....

[Article contains additional citation context not shown here]

P. Muth and T. C. Rakow. Atomic commitment for integrated database systems. In Proc. of the 7th Int'l Conf. on Data Engineering, 1991.

No context found.

P. Muth, Thomas C. Rakow: "Atomic Commitment for Integrated Database Systems", Proceedings of IEEE Seventh International Conference on Data Engineering, Kobe, Japan, April 8-12, 1991. Los Alamitos: IEEE 1991.

....(COM) The latter one uses external media servers for storage. This architecture enables easy integration of specific hardware as real time servers, tape or magneto optical jukeboxes or CD ROM devices. Here, a mechanism for atomic commitment of the resulting distributed transactions is required [6]. The Continuous Transport Manager (CTM) enables access to Internet protocols as well as Asynchronous Transport Management (ATM) The Spatial, Temporal and Interaction Script Interpreter (STI) interprets the time line based scripts (as generated from a VML schema) and triggers the Multimedia ....

P. Muth and T.C. Rakow. Atomic Commitment for Integrated Database Systems. In Proc. 7th Int. Conf. on Data Engineering (ICDE), pages 296--304, Feb. 1991.

.... Get (i1.QOH) Put (i1.QOH) ChangeStatus (o1, paid) Get (o1.Status) Put (o1.Status) PayOrder (i2,o2) ChangeStatus (o2, paid) Get (o2.Status) Put (o2.Status) ShipOrder (i2,o2) Put (o2.Status) Get (i2.QOH) Put (i2.QOH) tions and the special case of multilevel transactions are further discussed in [BSW88, Ga83, HW91, LKS91, MGG86, MR91, WHBM90, Wei91]; in the extended OODBS context of this paper, recovery issues are not yet addressed. It seems fairly natural to view the method invocation hierarchy of an OODBS transaction as an open nested transaction (cf. also [BM91, DHL91, RGN90] A concurrent execution of two transactions from the ....

Muth, P., Rakow, T., Atomic Commitment for Integrated Database Systems, IEEE Conf. on Data Engineering, 1991

....VODAK communication manager, the only additional requirement for implementing a heterogeneous three level transaction management is an atomic commitment protocol. Because we use inverse actions for recovery, there is no need to require the integrated systems to support a kind of two phase commit [MR91] We require only a function to detect the outcome of an L 0 transaction after a system crash. By using open nested and heterogeneous three level transactions, VODAK provides users with full fledged transaction support for transactions that spawn several integrated systems. Both criteria, high ....

P. Muth and T. Rakow. Atomic Commitment for Integrated Database Systems. In Proceedings of IEEE Seventh International Conference on Data Engineering, Kobe, Japan, 1991, pp. 296--304.

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