A formal characterization of SI-based ROWA replication protocols

Snapshot isolation (SI) is commonly used in some commercial DBMSs with a multiversion concurrency control mechanism since it never blocks read-only transactions. Recent database replication protocols have been designed using SI replicas where transactions are firstly executed in a delegate replica and their updates (if any) are propagated to the rest of the replicas at commit time; i.e. they follow the Read One Write All (ROWA) approach. This paper provides a formalization that shows the correctness of abstract protocols which cover these replication proposals. These abstract protocols differ in the properties demanded for achieving a global SI level and those needed for its generalized SI (GSI) variant — allowing reads from old snapshots. Additionally, we propose two more relaxed properties that also ensure a global GSI level. Thus, some applications can further optimize their performance in a replicated system while obtaining GSI.     

Transaction recovery in federated autonomous databases

Transaction management for federated database systems (FDBSs), where each participating DBMS is autonomous, supports global transactions that can access more than one database. A number of concurrency control algorithms exist for the failure-free environment. Handling transaction failure, due to concurrency control or site related reasons, becomes especially difficult in the presence of autonomy. Due to local autonomy, carrying out 2-phase commit may not be possible. This can be simulated by providing a server on top of the pre-existing DBMS at each site, which is responsible for submitting the local operations to the associated DBMS and communicating with the transaction's originating site. In this paper we formalize the problem of ensuring transaction consistency in an FDBS environment in the presence of failure. The key problem is that due to autonomy, the local DBMS and FDBS may have different views of an execution sequence generated at a site.Local recoverability is identified as the property of local execution sequences necessary for correctness. The other main problem is of guaranteeing that the various locally recoverable histories are consistent with each other. These identified properties are necessary and sufficient conditions for the correctness of FDBS recovery algorithms, and can be used to evaluate the correctness of the proposed algorithms. This paper also presents an FDBS recovery algorithm that has been proved to be correct. Formal proofs of all properties and a comparison of different algorithms are provided. Recommended by: Y. BreitbartThis work was supported in part by sub-contract #B09530013 from Honeywell SSDC, under contract #F30602-91-C-0128 from Rome Laboratory of the US Air Force.On leave from Computer Science Department, Heilongjiang University, People's Republic of China.     

Coordinator log transaction execution protocol

The coordinator log transaction execution protocol proposed in this paper centralizes logging on a per transaction basis and exploits piggybacking to provide the semantics of a distributed atomic commit at a minimal cost. The protocol eliminates two rounds of messages (one phase) from the presumed commit protocol and dramatically reduces the number of log forces needed for distributed atomic commit. We compare the coordinator log transaction execution protocol to existing protocols, explain when it is desirable, and discuss how it affects the write ahead log protocol and the database crash recovery algorithm. Recommended by: Tamer Ozsu     

移动分布式数据库中的实时事务处理

提出了移动事务实时提交协议(MTRTC)和多版本乐观并发控制协议(MVOCC)处理移动分布式实时事务。MVOCC有效性检查分为局部有效性检查和全局有效性检查,采用动态调整串行次序,避免了不必要的事务重启动,改善了只读事务的响应时间。MTRTC是一个实时提交协议,减少了通信信息。实验结果表明结合MVOCC和MTRTC事务处理协议优于其它事务处理协议。     

Divergence control for distributed database systems

This paper presents distributed divergence control algorithms for epsilon serializability for both homogeneous and heterogeneous distributed databases. Epsilon serializability allows for more concurrency by permitting non-serializable interleavings of database operations among epsilon transactions. We first present a strict 2-phase locking divergence control algorithm and an optimistic divergence control algorithm for a homogeneous distributed database system, where the local orderings of all the sub-transactions of a distributed epsilon transaction are the same. In such an environment, the total inconsistency of a distributed epsilon transaction is simply the sum of those of all its sub-transactions. We then describe a divergence control algorithm for a heterogeneous distributed database system, where the local orderings of all the sub-transactions of a distributed epsilon transaction may not be the same and the total inconsistency of a distributed epsilon transaction may be greater than the sum of those of all its sub-transactions. As a result, in addition to executing a local divergence control algorithm in each site to maintain the local inconsistency, a global mechanism is needed to take into account the additional inconsistency Recommended by: Meichum Hsu     

Transaction commit in a realistic timing model

Animportant problem in the construction of fault-tolerant distributed database systems is the design of nonblocking transaction commit protocols. This problem has been extensively studied for synchronous systems (i.e., systems where no messages ever arrive late). In this paper, the synchrony assumption is relaxed. A new partially synchronous timing model is described. Developed for this model is a new nonblocking randomized transaction commit protocol, which incorporates an agreement protocol of Ben-Or. The new protocol works as long as fewer than half the processors fail. A matching lower bound is proved, showing that the number of processor faults tolerated is optimal. If half or more of the processors fail, the protocol degrades gracefully: it blocks, but no processor produces a wrong answer. A notion of asynchronous round is defined, and the protocol is shown to terminate in a small constant expected number of asynchronous rounds. In contrast it is shown that no protocol in this model can guarantee that a processor terminates in a bounded expected number of its own steps, even if processors are synchronous. Brian A. Coan received the B.S.E. degree in electrical engineering and computer science from Princeton University, Princeton, New Jersey, in 1977; the M.S. degree in computer engineering from Stanford University, Stanford, California, in 1979; and the Ph.D. degree in computer science from the Massachusetts Institute of Technology, Cambridge, Massachusetts, in 1987. He has worked for Amdahl Corporation and AT & T Bell Laboratories. Currently he is a member of the technical staff at Bellcore. His main research interest is fault tolerance in distributed systems. Jennifer Lundelius Welch received her B.A. in 1979 from the University of Texas at Austin, and her S.M. and Ph.D. from the Massachusetts Institute of Technology in 1984 and 1988 respecively. She was a member of technical staff at GTE Laboratories Incorporated in Waltham, Massachusetts, from 1988 to 1989. She is currently an assistant professor at the University of North Carolina in Chapel Hill. Her research interests include algorithms and lower bounds for distributed computing.The authors were with the MIT Laboratory for Computer Science when the bulk of this work was done. This work was supported in part by the Advanced Research Projects Agency of the Department of Defense under Contract N00014-83-K-0125, the National Science Foundation under Grant DCR-83-02391, the Office of Army Research under Contract DAAG29-84-K-0058, and the Office of Naval Research under Contract N00014-85-K-0168. A preliminary version of this paper appears in theProceedings of the Fifth Annual ACM Symposium on Principles of Distributed Computing [2]     

Highly available and reliable communication protocols for heterogeneous systems

Existing and legacy software systems are the product of lengthy and individual development histories. Interoperability among such systems offers the support of global applications on these systems. However, interoperability among these heterogeneous systems is hampered by the absence of a reliable communication environment that supports the development of global applications.In this paper, we show how a generic communication framework can serve as a testbed for the specification, verification, and execution of distributed communication protocols. The development of distributed, global concurrency protocols is much simpler than using traditional tools, like RPC (remote procedure call), because our framework provides a high-level communication mechanism that frees the protocol designer from thinking in a message-based style. We present several protocols that are consistent with realistic assumptions about local database systems, and proofs of their correctness and consistency preservation. We also show that the execution of these protocols is fault-tolerant. The distribution of systems can be chosen according to application requirements, without adaptation of protocols. Fault tolerance can be fine-tuned within the framework itself, so that verified protocols do not need modifications in this case either.Besides protocols for traditional transaction processing, we present communication protocols for advanced transaction models that relax one or more of the ACID properties of transactions. These advanced transaction models enable local autonomy and thus are much better suited for heterogeneous environments.     

Transaction management issues in a failure-prone multidatabase system environment

This paper is concerned with the problem of integrating a number of existing off-the-shelf local database systems into a multidatabase system that maintains consistency in the face of concurrency and failures.The major difficulties in designing such systems stem from the requirements that local transactions be allowed to execute outside the multidatabase system control, and that the various local database systems cannot participate in the execution of a global commit protocol. A scheme based on the assumption that the component local database systems use the strict two-phase locking protocol is developed. Two major problems are addressed: How to ensure global transaction atomicity without the provision of a commit protocol, and how to ensure freedom from global deadlocks.     

实时Client/Server数据库并发控制和恢复机制研究

该文提出了实时Client／Server数据库系统多版本两阶段封锁并发控制协议和有效的恢复机制。协议区分只读事务和更新事务。只读事务在执行读操作时遵从多版本时间排序协议,更新事务执行强两阶段封锁协议,即持有全部锁直到事务结束。只读事务读请求从不失败,不必等待等特性。在典型数据库系统中,读操作比写操作频繁。这个特性对于实践来说至关重要。为了提高只读事务的响应时间,协议让每个客户端与一个一致数据库影子相联,只读事务在客户端处理。更新事务提交到服务端运行。服务端每个事务Ti在提交时系统必须向所有客户端广播信息。客户端根据得到的广播信息自动构造一致数据库影子。一致数据库影子还将用于系统恢复。通过仿真模拟。与2V2PL和OCC-TI-WAIT-50协议进行比较,结果表明：该并发控制协议不仅能有效降低事务延误截止时间率和重起动率,而且能改善只读事务的响应时间,减少优先级高事务的锁等待时间。协议性能优于2V2PL协议和OCC-TI-WAIT-50协议。     

Overview of multidatabase transaction management

A multidatabase system (MDBS) is a facility that allows users access to data located in multiple autonomous database management systems (DBMSs). In such a system,global transactions are executed under the control of the MDBS. Independently,local transactions are executed under the control of the local DBMSs. Each local DBMS integrated by the MDBS may employ a different transaction management scheme. In addition, each local DBMS has complete control over all transactions (global and local) executing at its site, including the ability to abort at any point any of the transactions executing at its site. Typically, no design or internal DBMS structure changes are allowed in order to accommodate the MDBS. Furthermore, the local DBMSs may not be aware of each other and, as a consequence, cannot coordinate their actions. Thus, traditional techniques for ensuring transaction atomicity and consistency in homogeneous distributed database systems may not be appropriate for an MDBS environment. The objective of this article is to provide a brief review of the most current work in the area of multidatabase transaction management. We first define the problem and argue that the multidatabase research will become increasingly important in the coming years. We then outline basic research issues in multidatabase transaction management and review recent results in the area. We conclude with a discussion of open problems and practical implications of this research.     

A toolkit for the incremental implementation of heterogeneous database management systems

The integration of heterogeneous database environments is a difficult and complex task. The A la carte Framework addresses this complexity by providing a reusable and extensible architecture in which a set of heterogeneous database management systems can be integrated. The goal is to support incremental integration of existing database facilities into heterogeneous, interoperative, distributed systems. The Framework addresses the three main issues in heterogeneous systems in tegration. First, it identifies the problems in integrating heterogeneous systems. Second, it identifies the key interfaces and parameters required for autonomous systems to interoperate correctly. Third, it demonstrates an approach to integrating these interfaces in an extensible and incremental way. The A la carte Framework provides a set of reusable, integrating components which integrate the major functional domains, such as transaction management, that could or should be integrated in heterogeneous systems. It also provides a mechanism for capturing key characteristics of the components and constraints which describe how the components can be mixed and interchanged, thereby helping to reduce the complexity of the integration process. Using this framework, we have implemented an experimental, heterogeneous configuration as part of the object management work in the software engineering research consortium, Arcadia.     

多数据库事务处理模型中局部代理的设计与实现

在多数据库事务处理模型中，局部代理提供了全局事务管理层与局部数据库系统的接口，它使多数据库系统能够对分布在不同站点的局部数据库进行透明访问。文中提出了基于两阶段代理（２ＰＣＡ）的局部代理模型，它集成了支持两阶段提交或单阶段提并的数据库系统、以及并不支持事务提交协议的文件系统。文中还介绍了采用专用接口和ＯＤＢＣ两种实现局部代理的方法。文中着重介绍了使用ＯＤＢＣ方法的具体实现，并对两种方法实现的局部代     

The Performance of Two Phase Commit Protocols in the Presence of Site Failures

The two phase commit is an important protocol in distributed database systems. Much of the existing literature on the protocol is restricted to discussing and analyzing the protocol (and its variants) in the absence of failures. Very little, especially in quantitative terms, has been written about its performance in the presence of site failures. In this study, we use a simulation testbed of a distributed database system to quantify the differences in the performances of four widely known variants of the 2PC protocols (the generic 2PC, presumed commit, presumed abort, and early prepare). Our study covers both the no-failure case and the case of site failures. We present a number of interesting results based on our experiments. One is that the performance of these protocols is highly dependent on the message-processing latency at the transaction coordinator site. Another is that the presumed abort protocol does not necessarily yield better performance in the presence of site failures.     

The PROMPT real-time commit protocol

We investigate the performance implications of providing transaction atomicity for firm-deadline real-time applications, operating on distributed data. Using a detailed simulation model, the real-time performance of a representative set of classical transaction commit protocols is evaluated. The experimental results show that data distribution has a significant influence on real-time performance and that the choice of commit protocol clearly affects the magnitude of this influence. We also propose and evaluate a new commit protocol, PROMPT (Permits Reading Of Modified Prepared-data for Timeliness), that is specifically designed for the real-time domain. PROMPT allows transactions to “optimistically” borrow, in a controlled manner, the updated data of transactions currently in their commit phase. This controlled borrowing reduces the data inaccessibility and the priority inversion that is inherent in distributed real-time commit processing. A simulation-based evaluation shows PROMPT to be highly successful, as compared to the classical commit protocols, in minimizing the number of missed transaction deadlines. In fact, its performance is close to the best on-line performance that could be achieved using the optimistic lending approach. Further, it is easy to implement and incorporate in current database system software. Finally, PROMPT is compared against an alternative priority inheritance-based approach to addressing priority inversion during commit processing. The results indicate that priority inheritance does not provide tangible performance benefits     

Transaction management in distributed heterogeneous database management systems

Current distributed, heterogeneous database management systems (DBMSs) address the issue of distributed transaction management (DTM) in two different ways: some systems rely solely on unmodified local transaction managers (LTMs), thereby helping preserve local DBMS autonomy, but limit functionality (e.g. allow only unsynchronized retrievals, preclude distributed updates, etc.); others maintain full functionality but require the (re)design of the LTMs to enforce homogeneous DTM across all heterogeneous DBMSs, thereby giving up most of the local DBMS autonomy.

The goal of the work presented here is to establish the minimum set of modifications to LTMs that allow synchronized retrievals and distributed updates (whenever semantic conflicts can be resolved), and will continue to maintain a high degree of local DBMS autonomy. The problems of: (1) distribution of responsibility between DTM and LTMs, (2) concatenation of functionally-equivalent LTM mechanisms, and (3) providing compensation mechanisms for functionally-limited LTMs are introduced. Solutions to the above problems are shown to exist. However, the interconnection of autonomous, heterogeneous DBMSs is significantly more difficult than anticipated, despite communication standardization and current optimism.     

An extended transaction to maintain consistency and recovery in multidatabase systems

A complete practical solution to transaction management preserving multidatabase consistency in the presence of multidatabase updates and failures is presented. The approach developed does not require support for the two-phase commit (2PC) protocol in the participating local database management systems (LDBMSs). Furthermore, it does not violate local autonomy; the source code of the LDBMSs is not modified in any way and the multidatabase system (MDBS) does not access or modify any control information of the LDBMS. The principles of the 2PC protocol in the process of global transaction commitment are adopted. The presented method does not rely on any specific concurrency control mechanism for LDBMSs. Consideration is given to global transaction failures due to subtransaction aborts by the LDBMSs and local site crashes. The recovery process is based on undo operations. While a global transaction is in progress, the tables accessed by subtransactions of this transaction at each local site are locked using specially initiated table locks. These locks are stored and maintained in the local database itself as control tables. The approach taken is easy to implement, and its limitations are discussed.     

Global lock escalation in database management systems

Since database management systems (DBMSs) have limited lock resources, transactions requesting locks beyond the limit must be aborted, degrading the performance abruptly. Lock escalation can be effectively used in such circumstances to alleviate the problem. Many lock escalation methods have been proposed and implemented in commercial DBMSs. However, they have certain problems due to the local nature of their decisions on when to execute lock escalation. In this paper, we propose a new lock escalation method, global lock escalation, that makes such decision globally based on the total number of locks. Through extensive simulation, we show that the global lock escalation method outperforms the existing ones significantly. Especially, we show that the number of concurrent transactions allowable increases by 2-16 times. We believe our method can be easily implemented in the commercial DBMSs enhancing the performance significantly under excessive lock requests.     

Overcoming Heterogeneity and Autonomy in Multidatabase Systems

A multidatabase system (MDBS) is a software system for integration of preexisting and independent local database management systems (DBMSs). The transaction management problem in MDBSs consists of designing appropriate software, on top of local DBMSs, such that users can execute transactions that span multiple local DBMSs without jeopardizing database consistency. The difficulty in transaction management in MDBSs arises due to the heterogeneity of the transaction management algorithms used by the local DBMSs, and the desire to preserve their local autonomy. In this paper, we develop a framework for designing fault-tolerant transaction management algorithms for MDBS environments that effectively overcomes the heterogeneity- and autonomy-induced problems. The developed framework builds on our previous work. It uses the approach described in S. Mehrotra et al. (1992, in “Proceedings of ACM–SIGMOD 1992 International Conference on Management of Data, San Diego, CA”) to overcome the problems in ensuring serializability that arise due to heterogeneity of the local concurrency control protocols. Furthermore, it uses a redo approach to recovery for ensuring transaction atomicity (Y. Breitbart et al., 1990, in “Proceedings of ACM–SIGMOD 1990 International Conference on Management of Data, Atlantic City, NJ;” Mehrotra et al., 1992, in “Proceedings of the Eleventh ACM SIGACT–SIGMOD–SIGART Symposium on Principles of Database Systems, San Diego, CA;” and A. Wolski and J. Veijalainen, 1990, in “Proceedings of the International Conference on Databases, Parallel Architectures and Their Applications”, pp. 321–330), that strives to ensure atomicity of transactions without the usage of the 2PC protocol. We reduce the task of ensuring serializability in MDBSs in the presence of failures to solving three independent subproblems, solutions to which together constitute a complete strategy for failure-resilient transaction management in MDBS environments. We develop mechanisms with which each of the three subproblems can be solved without requiring any changes be made to the preexisting software of the local DBMSs and without compromising their autonomy.     

Independent updates and incremental agreement in replicated databases

Update propagation and transaction atomicity are major obstacles to the development of replicated databases. Many practical applications, such as automated teller machine networks, flight reservation, and part inventory control, do not require these properties. In this paper we present an approach for incrementally updating a distributed, replicated database without requiring multi-site atomic commit protocols. We prove that the mechanism is correct, as it asymptotically performs all the updates on all the copies. Our approach has two important characteristics: it is progressive, and non-blocking.Progressive means that the transaction's coordinator always commits, possibly together with a group of other sites. The update is later propagated asynchronously to the remaining sites.Non-blocking means that each site can take unilateral decisions at each step of the algorithm. Sites which cannot commit updates are brought to the same final state by means of areconciliation mechanism. This mechanism uses the history logs, which are stored locally at each site, to bring sites to agreement. It requires a small auxiliary data structure, called reception vector, to keep track of the time unto which the other sites are guaranteed to be up-to-date. Several optimizations to the basic mechanism are also discussed. Recommended by: Ahmed Elmagarmid