基于代理的3PC分布式事务提交协议

原有3PC分布式事务提交协议能克服协同者发生故障而有可能产生事务阻塞问题,但其开销大。该文提出了一种基于代理的3PC事务提交协议,该协议通过增加协同者的代理节点,使得参与者相信协同者是正常的,从而不必关心新协同者的选举问题,能降低3PC无故障时的额外开销。     

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.     

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     

一种改进的基于多站点备份的1-2PC协议

2PC协议具有广泛的适用性,但存在易于阻塞和时间延迟的缺点。文中提出了一种改进的基于多站点备份的1-2PC协议。该协议利用＂多站点备份＂技术和＂心跳＂技术可减少分布式事务信息交换的次数,降低事务发生阻塞的概率,缩短事务成功提交的时间,从而改善了提交协议的性能。通过性能比较和分析,证明了提出的协议具有高效性和低阻塞性。     

移动实时提交协议的恢复处理与通信分析

三层实时提交协议(3LRTC)是为支持事务移动性、断接性、实时性及功能替代性等设计的一种移动实时提交协议,该协议能保证移动分布式实时事务放松原子性和结构正确性。该文在分析移动实时事务各参与者提交状态变化的基础上,讨论了3LRTC在发生场地故障和断接时的恢复处理方式,以保证故障发生后数据库能恢复到一个正确的状态。该文还对3LRTC的通信代价进行了分析与比较,结果表明,三层提交结构能减少移动实时事务的通信消息数。     

一个带有时限的工程设计事务提交协议

通过分析协作设计活动的特点,给出了一个协作设计事务模型,并在此基础上提出了一个带有时限的工程设计事务两阶段提交协议,包括协议描述及故障处理。分析表明,带有时限的工程设计事务提交协议的性能在所有情况下优于基本两阶段提交协议。     

Dictatorial Transaction Processing: Atomic Commitment Without Veto Right

The current standard in governing distributed transaction termination is the so-called Two-Phase Commit protocol (2PC). The first phase of 2PC is a voting phase, where the participants in the transaction are given an ultimate right to abort that transaction. Giving up that veto right from all participants reduces the overhead of the atomic commitment protocol but also imposes some restrictions on the concurrency control and recovery protocols employed by the participants in the transaction.This paper gives, for the first time, a precise abstract specification of the Dictatorial Atomic Commitment (DAC) problem, resulting from removing veto rights from the traditional Atomic Commitment (AC) problem. We characterize transactional systems that are compatible with that specification in terms of necessary and sufficient conditions on concurrency control and recovery protocols, and discuss the practical impacts of those conditions. From this study, we capitalize on existing protocols that solve the DAC problem, and propose a new protocol that broadens the applicability of dictatorial transaction processing in order to meet the requirements of today's distributed environments. We point out interesting performance tradeoffs, and describe the implementation of our protocol in the context of current transactional standards, initially designed with 2PC in mind.     

Fast non-blocking atomic commit: an inherent trade-off

This paper investigates the time-complexity of the non-blocking atomic commit (NBAC) problem in a synchronous distributed model where t out of n processes may fail by crashing. We exhibit for t?3 an inherent trade-off between the fast abort property of NBAC, i.e., aborting a transaction as soon as possible if some process votes “no”, and the fast commit property, i.e., committing a transaction as soon as possible when all processes vote “yes” and no process crashes. We also give two algorithms: the first satisfies fast commit and a weak variant of fast abort, whereas the second satisfies fast abort and a weak variant of fast commit.     

数据访问的事务处理中间件中等待时间的研究

数据访问的事务处理中间件主要用于分布式的数据库访问的事务处理，采用两阶段提交技术。在两阶段提交中当协调者把任务分支分配到所有的参与者后，需要在等待一定的时间再发出准备消息。等待时间的大小会对分布式系统的并发性产生影响，为了提高系统的并发性，需要对提交等待时间进行估计。该论文对于数据访问的事务处理中间件任务的类型较少的特点，提出了一种对于提交等待时间进行较准确估计的方法，且针对不同类型的数据访问(单个操作和批量操作)采用不同的时间估计，提出了长事务的概念。     

移动数据库事务处理模型的研究

在移动数据库系统中,事务的移动性、频繁断接性以及长事务等特性使得传统的事务处理模式不再适用。提出一种扩展的乐观两阶段提交事务处理模型（O2PC-MT）,该模型吸收了O2PC-MT模型的设计思想,解决了由短暂的通信失效造成不必要的事务中止以及移动事务协调器故障导致的阻塞等问题。实验结果表明,与O2PC-MT事务处理模型相比,EO2PC-MT提高了系统的事务吞吐率以及改善了系统的总体性能。     

基于环境一致性的分布式并行数据更新协议

为了保证服务器节点完全对等的分布式并行数据库系统更新结果一致性,提出基于数据环境一致性的分布式并行更新协议DEC-DP2PC(Distributed and Parallel 2PC Based on Data-Environment-Consistency),对经典2PC(two phase commit)在执行条件、表决内容和次数、全局提交和撤销定义等方面进行了针对性改进。性能分析和测试表明,DEC-DP2PC可有效过滤数据副本分布和版本号等数据环境信息不一致情况下事务的执行,减少全局数据环境一致性维护开销和降低系统平均响应时间。     

Commit protocols for externalized-commit heterogeneous database systems

In designing a heterogeneous database systems, one of the main technical challenges is developing techniques for ensuring global commit. That is, guaranteeing that a transaction spanning multiple individual database management systems (DBMSs) either commits at all the participating DBMSs or at none of them. Previous work in this area typically assumes that the participating DBMSs do not provide a mechanism for interacting with their commit facilities. While this is true in many cases, in practice there are systems which support a programmatic interface to their commit protocols. We refer to database systems offering such facilities asexternalized commit DBMSs.The focus of this paper is on commit protocols for these systems. We propose two new commit protocols for externalized commit DBMSs. The first may be used to obtain global commit in heterogeneous database systems composed of DBMSs with different 2-phase commit protocols (e.g., centralized and linear). The second protocol is more general, and ensures global commit even if the participating DBMSs employ 3-phase commit protocols. The more general protocol also preserves database autonomy, since it does not block a DBMS upon failure of another system. We describe both protocols in detail and prove their correctness. Recommended by: M. RusinkiewiczThis work was partially supported by an IBM Research Initiation Grant.     

A cross-layer atomic commit protocol implementation for transaction processing in mobile ad-hoc networks

Transaction processing leads to new challenges in mobile ad-hoc networks, which, in comparison to fixed-wired networks, suffer from problems like node disconnection, message loss, and frequently appearing network partitioning. As the atomic commit protocol is that part of transaction processing in which failures can lead to the most serious data blocking, we have developed a robust and failure-tolerant distributed cross-layer atomic commit protocol called CLCP that uses multiple coordinators. In order to reduce the number of both, failures and messages, our protocol makes use of acknowledgement messages for piggybacking information. We have evaluated our protocol in mobile ad-hoc networks by using several mobility models (i.e. Random Waypoint, Manhattan, and Attraction Point), and compared CLCP with other atomic commit protocols, i.e. 2PC and Paxos Commit, each implemented in 3 versions, i.e. without sending message acknowledgements, with a Relay Routing technique, and with Nearest Forward Progress Routing. Special to our simulation environment is the use of the quasi-unit-disc model, which assumes a non-binary message reception probability that captures real-world behavior much better than the classical unit-disc-model, often used in theory. Using the quasi-unit-disc model, our evaluation shows the following results. CLCP and “2PC without acknowledgement messages” have a significantly lower energy consumption than the other protocols, and CLCP is able to commit significantly more distributed transactions than all the other atomic commit protocols for each of the mobility models.     

分布式实时事务提交协议

在分布式实时数据库系统中，保证事务原子性的唯一途径是研究和开发出一个实时的原子提交协议．首先详细分析了事务因数据访问冲突而形成的各种依赖关系，在此基础上提出了实时的原子乐观提交协议——2SC协议，该协议减少了事务的等待时间，提高了事务的并发度，且能无缝地和现有的并发控制协议集成在一起，保证事务的可串行化和原子性．通过模拟实验研究表明，采用该协议能够减少超过截止期的事务数目。     

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     

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]     

分布式实时事务提交处理

由于提交处理的复杂性,分布式实时事务很难满足其截止期.提出了一种新的提交协议A2SC(主动的双空间提交),它适合于分布式实时事务提交处理的需要.分析了由于数据冲突访问而形成的各种依赖关系.当处于准备状态的事务和处于提交状态的事务发生数据冲突访问时,A2SC允许处于执行状态的事务在一种控制的方式下乐观地访问锁住的数据.当处于准备状态的事务夭折时,仅仅只有其夭折依赖集中的事务夭折.进一步提出了"没有结果的运行"的观念.当一个事务发现它是没有结果的允许时,它将主动夭折.进行了广泛的模拟实验比较A2SC和其它协议比如基准协议、PROMPT和DDCR的性能.模拟结果表明A2SC在最小化错过截止期的事务数方面较成功,因此A2SC适合于高性能分布式实时事务.     

Blocking reduction for distributed transaction processing within MANETs

Atomic commit protocols for distributed transactions in mobile ad-hoc networks have to consider message delays and network failures. We consider ad-hoc network scenarios, in which participants hold embedded databases and offer services to other participants. Services that are composed of several other services can access and manipulate data of physically different databases. In such a scenario, distributed transaction processing can be used to guarantee atomicity and serializability throughout all databases. However, with problems like message loss, node failure, and network partitioning, mobile environments make it hard to get estimations on the duration of a simple message exchange. In this article, we focus on the problem of setting up reasonable time-outs when guaranteeing atomicity for transaction processing within mobile ad-hoc networks, and we show the effect of setting up “wrong” time-outs on the transaction throughput and blocking time. Our solution, which does not depend on time-outs, shows a better performance in unreliable networks and remarkably reduces the amount of blocking.     

移动分布式实时嵌套事务提交

在移动分布式计算环境中,事务移动性和无线网络固有的缺陷使得传统的分布式实时事务管理机制不足以支持移动分布式实时事务的执行,故有必要为移动实时事务研究新的事务处理机制,以提高其成功率.着重研究移动实时事务的提交机制.首先,通过分析移动分布环境中实时事务的特点给出了一个基于功能替代的移动实时嵌套事务模型.然后,提出了一个基于此模型的三层提交结构以及能够保证移动实时事务原子性和结构正确性的三阶段实时提交协议3PRTC(three-phase real-time commit).性能测试表明,所提出的事务模型及其提交机制能够提高实时事务的成功率.