Optimistic Concurrency Control Scheme Research Articles

A survey on optimizations towards best-effort hardware transactional memory

Transactional memory has been attracting increasing attention in recent years, and it provides optimistic concurrency control schemes for shared-memory parallel programs. The rapid development and wide adoption of transactional memory make this programming paradigm promising for achieving breakthroughs in massively parallel computing. There has been a large number of discussions towards transactional memory systems, which aimed at providing relatively simple and intuitive synchronization construction for shared-memory parallel programs without sacrificing performance. Hardware transactional memory (HTM) has become commercially available in mainstream processors, however, due to several inherent architectural limitations that will abort hardware transactions, such as cache overflows, context switches, hardware as well as software exceptions, etc., nowadays HTM systems come in a best-effort way, which necessitates the adoption of a software fallback path to ensure forward progress. In this paper, we survey state-of-the-art software-side optimizations for best-effort hardware transaction system, as well as several novel performance tuning techniques. Research efforts about joint usage of HTM and non-volatile memory (NVM) are also discussed.

Relaxing real-time order in opacity and linearizability

In this paper we introduce two families of safety properties: ♢-opacity and ♢-linearizability. The new properties relax (to a various degree) the real-time order requirement on transaction execution in opacity and, analogically, the real-time order requirement on operation execution in linearizability. This way we can formalize the guarantees provided by a wide class of strongly consistent replicated systems for which opacity and linearizability are too strong. We show the formal relationship between ♢-opacity and ♢-linearizability which allows us to directly compare semantics of transactional and non-transactional systems and, in particular, opacity and linearizability in their original definitions. We also illustrate how the new properties can be used by proving correctness of Deferred Update Replication, a well known optimistic concurrency control scheme. We show that it satisfies update-real-time opacity, a member of the ♢-opacity family, which allows read-only and aborted transactions to operate on stale (but still consistent) data.

Towards elastic transactional cloud storage with range query support

Cloud storage is an emerging infrastructure that offers Platforms as a Service (PaaS). On such platforms, storage and compute power are adjusted dynamically, and therefore it is important to build a highly scalable and reliable storage that can elastically scale on-demand with minimal startup cost. In this paper, we propose ecStore -- an elastic cloud storage system that supports automated data partitioning and replication, load balancing, efficient range query, and transactional access. In ecStore, data objects are distributed and replicated in a cluster of commodity computer nodes located in the cloud. Users can access data via transactions which bundle read and write operations on multiple data items stored on possibly different cluster nodes. The architecture of ecStore follows a stratum design that leverages an underlying distributed index with a replication layer in the middle and a transaction management layer on top. ecStore provides adaptive read consistency on replicated data. We also enhance the system with an effective load balancing scheme using a self-tuning replication technique that is specially designed for large-scale data. Furthermore, a multi-version optimistic concurrency control scheme matches well with the characteristics of data in cloud storages. To validate the performance of the system, we have conducted extensive experiments on various platforms including a commercial cloud (Amazon's EC2), an in-house cluster, and PlanetLab.

A hybrid distributed optimistic concurrency control method for high-performance real-time transaction processing

The conventional lock scheme tends to suffer from a cascade of blockings, while the optimistic concurrency control (OCC) scheme may suffer from wasting resources. To overcome these problems, some researchers have proposed a combination of OCC and lock in transaction processing. Using this method, Thomasian proposed the hybrid method for conventional distributed transaction processing, and Lam proposed the DOCC-DA protocol for distributed real-time database system based on forward validation. This paper proposes a new protocol, called Hybrid Distributed Optimistic Concurrency Control Embedded in two-Phase Commit, which is based on back validation. The new protocol makes use of access invariance and runtime information which can guarantee a rerun transaction to meet its deadline and abort the fruitless run transactions as early as possible. A series of simulation experiments have been done to investigate the performance of the new protocol. The results show that its performance is consistently better than that of other protocols.

Read-down conflict-preserving serializability as a correctness criterion for multilevel-secure optimistic concurrency control: CSR/RD

It is important to note that conflict-preserving serializability and related theory are artifacts of an era of database development where correctness alone was the overriding concern. With the advent of multilevel-secure databases, there is clearly a need to reexamine such theories. Any correctness criterion to govern transaction processing in the multilevel security context has to incorporate both secureness and correctness in a unified manner. This paper makes original contributions in two different but closely related areas to the optimistic concurrency control in multilevel-secure, single-version databases. First, read-down conflict-preserving serializability (CSR/RD) captures multilevel-secure database consistency requirements and secure transaction correctness properties via a single notion. Second, it presents a multilevel-secure optimistic concurrency control (MLS/OCC) scheme that has several desirable properties: If lower-level transactions were somehow allowed to continue with their executions in spite of the conflict with high-level transactions, covert timing-channel freeness would be satisfied. This sort of optimistic approach for conflict insensitiveness and the properties of non-blocking and deadlock freedom make the optimistic concurrency control scheme especially attractive to multilevel-secure transaction processing.

An adaptable constrained locking protocol for high data contention environments: correctness and performance

Multiversions of data are used in database systems to increase concurrency and to provide efficient recovery. Data versions improve the concurrency by allowing the concurrent execution of “non-conflicting” read-write lock requests on different versions of data in an arbitrary fashion. A transaction that accesses a data item version, which later diagnosed to lead to an incorrect execution, is aborted. This act is reminiscent of the validation phase in the optimistic concurrency control schemes. Various performance studies suggest that these schemes perform poorly in high data contention environments where the excessive transaction aborts result due to the failed validation. We propose an adaptable constrained two-version two-phase locking (C2V2PL) scheme in which these “non-conflicting” requests are allowed only in a constrained manner. C2V2PL scheme assumes that a lock request failing to satisfy the specific constraints will lead to an incorrect execution and hence, must be either rejected or blocked. This eliminates the need for a separate validation phase. When the contention for data among the concurrent transactions is high, the C2V2PL scheduler in aggressive state rejects such lock requests. The deadlock free nature of C2V2PL scheduler adapts to the low data contention environments by accepting the lock request that have failed the specific constraints but contrary to the assumption, will not lead to an incorrect execution. Thus, C2V2PL scheme improves the potential concurrency due to reduced transaction aborts in this conservative state. We have compared performance of our scheme with other lock-based concurrency control schemes such as two phase locking, Wait-depth locking and Optimistic locking schemes. Our results show increase in throughput and reduced transaction-abort-ratio in case of C2V2PL scheme.

Efficient optimistic concurrency control using loosely synchronized clocks

This paper describes an efficient optimistic concurrency control scheme for use in distributed database systems in which objects are cached and manipulated at client machines while persistent storage and transactional support are provided by servers. The scheme provides both serializability and external consistency for committed transactions; it uses loosely synchronized clocks to achieve global serialization. It stores only a single version of each object, and avoids maintaining any concurrency control information on a per-object basis; instead, it tracks recent invalidations on a per-client basis, an approach that has low in-memory space overhead and no per-object disk overhead. In addition to its low space overheads, the scheme also performs well. The paper presents a simulation study that compares the scheme to adaptive callback locking, the best concurrency control scheme for client-server object-oriented database systems studied to date. The study shows that our scheme outperforms adaptive callback locking for low to moderate contention workloads, and scales better with the number of clients. For high contention workloads, optimism can result in a high abort rate; the scheme presented here is a first step toward a hybrid scheme that we expect to perform well across the full range of workloads.

Concurrency control for object bases

This paper presents concurrency control techniques suitable for engineering design environments. Design environments are quite different from conventional business applications for which databases have been traditionally used. Transactions in a design application are usually very long compared with those in traditional applications, necessitating new concurrency control techniques. We present the Dynamic Validation concurrency control scheme, an optimistic concurrency control scheme that avoids locking data items for extended periods of time. Dynamic validation is designed specifically for low-contention object-oriented database applications with long-duration transactions, such as engineering design applications. We prove that dynamic validation enforces executions that are “serializable with respect to the database state”. This correctness criterion relaxes the classical serializability criterion, and allows schedules that are equivalent, i.e. produce the same effect, to a serial execution of transactions starting from the same initial database state.

Performance analysis of concurrency control using locking with deferred blocking

The concurrency control (CC) method employed can be critical to the performance of transaction processing systems. Conventional locking suffers from the blocking phenomenon, where waiting transactions continue to hold locks and block other transactions from progressing. In a high data contention environment, as an increasing number of transactions wait, a larger number of lock requests get blocked and fewer lock requests can get through. The proposed scheme reduces the blocking probability by deferring the blocking behavior of transactions to the later stages of their execution. By properly balancing the blocking and abort effects, the proposed scheme can lead to better performance than either the conventional locking or the optimistic concurrency control (OCC) schemes at all data and resource contention levels. We consider both static and dynamic approaches to determine when to switch from the nonblocking phase to the blocking phase. An analytical model is developed to estimate the performance of this scheme and determine the optimal operating or switching point. The accuracy of the analytic model is validated through a detailed simulation. >

Analysis of hybrid concurrency control schemes for a high data contention environment

Analytical models are developed to study hybrid CC (concurrency control) schemes which employ a different CC scheme to handle rerun transactions, since their characteristics are different from the first run of transactions. These include switching to static or dynamic locking during rerun (referred to as static and dynamic hybrid OCC (optimistic concurrency control) schemes, respectively), and switching to broadcast OCC during rerun, while doing pure OCC for the first run. In a high data contention environment where locking is inferior to OCC, analysis shows that the performance can be substantially improved by using this hybrid approach and the authors study the tradeoff of the different hybrid CC schemes. The analytic models are based on a decomposition approach and use a mean-value-type analysis. The accuracy of the analysis is validated through simulations. >

Poisson winner queues

Kleinrock, L. and F. Mehović, Poisson winner queues, Performance Evaluation 14 (1992) 79–101. We study special types of queues, winner queues, in which all customers are served concurrently. A customer in a winner queue will successfully finish his service (i.e. “win”) and leave if no other customer leaves during his current service. Once a customer wins, all others in service at that time “lose”; thus we have a situation in which concurrent customers conflict, yielding a single “winner”. There are four disciplines considered: silent-redraw, silent-noredraw, broadcast-redraw, and broadcast-noredraw. The winner queues considered have an infinite number of servers. We assume that service times consist of a deterministic part and an exponential part. This type of service time distribution includes pure deterministic and pure exponential service times as special cases. Using a one-dimensional imbedded Markov chain and a recursive formula for the state probabilities, we obtain numerical results for certain disciplines and distributions of requested and restarted service time. For some broadcast winner queues we show analytic results. In all cases we also give simulation results which indicate the correctness and accuracy of our numerical calculations. The results obtained are given in terms of the normalized average system time and the normalized power (defined as the system load divided by the normalized average system time). One application of this model is to study the performance of optimistic concurrency control schemes in databases.

Concurrency control optimizations in a prolog database

The degree of concurrency allowed by a concurrency control scheme depends on the knowledge it has about the database and the transactions. This paper analyzes the syntactic information available in a Prolog database to improve concurrency. An optimistic concurrency control scheme is chosen since it is a natural choice for Prolog databases. The conflict criteria among transactions in a Prolog environment is translated into conditions on the query search tree. This aids in identifying concurrency related information from the depth-first search execution strategy. A spesented in stages such that progressively more information is considered to construct the read and the write sets (required to check conflicts) and better concurrency is achieved. An orthogonal contribution of the information analysis is a new query evaluation optimization. The price for better concurrency is paid in terms of increased storage and computation costs. The optimality of the algorithms with respect to the available information and derived concurrency is discussed qualitatively.

A space-efficient technique for recording versions of data

This paper describes a space-efficient technique for recording successive differences between versions of data. The technique is based on the combined use of differential files and change area b-trees. The strengths of the technique are that a branching tree of versions can be accommodated; any data item in any version can be accessed with approximately equal access time; and very large volumes of data can be accommodated. The paper also describes the use of this technique to support very long transactions in the SDS2 software development system. Here the technique is used in an optimistic concurrency control scheme.

Performance analysis of centralized databases with optimistic concurrency control

In this paper we analyze Optimistic Concurrency Control (OCC) schemes based on certification. We allow two data access schemes referred to as static and dynamic. A static data access scheme implies that all data items are accessed at the beginning of transaction processing, while data items are accessed on demand during the processing of a transaction in a dynamic scheme. After completing its processing, each transaction is validated, i.e., checked as to whether it has encountered a data conflict. Validated transactions commit; otherwise, they are restarted. A variant of the regular (silent) commit scheme where a committing transaction notifies conflicted transactions to restart immediately (broadcast commit scheme) is also considered. System efficiency or the fraction of useful work as affected by data contention is used as the main performance measure. We develop an iterative solution method to analyze the performance of OCC schemes in the framework of a system with a fixed number of transactions in multiple classes with given occurrence frequencies. The model is analyzed by (iteratively) solving a set of nonlinear equations which relate per class efficiencies to system efficiency. The iterative solution method is validated against simulation and shown to be highly accurate even for high data contention. An advantage of the iterative solution method is that it separates the analysis of data contention from hardware resource contention and provides per class efficiencies. We present graphs/tables, which are used to determine how system performance is affected by: (i) various OCC schemes, (ii) the distribution of transaction processing times, (iii) transaction size and its distribution, and (iv) the throughput characteristics of the system (as determined by hardware contention).

Distributed optimistic concurrency control with reduced rollback

Concurrency control algorithms have traditionally been based on locking and timestamp ordering mechanisms. Recently optimistic schemes have been proposed. In this paper a distributed, multi-version, optimistic concurrency control scheme is described which is particularly advantageous in a query-dominant environment. The drawbacks of the original optimistic concurrency control scheme, namely that inconsistent views may be seen by transactions (potentially causing unpredictable behavior) and that read-only transactions must be validated and may be rolled back, have been eliminated in the proposed algorithm. Read-only transactions execute in a completely asynchronous fashion and are therefore processed with very little overhead. Furthermore, the probability that read-write transactions are rolled back has been reduced by generalizing the validation algorithm. The effects of global transactions on local transaction processing are minimized. The algorithm is also free from dedlock and cascading rollback problems.

Analysis of some optimistic concurrency control schemes based on certification

Optimistic Concurrency Control-OCC schemes based on certification are analyzed in this paper. We allow two types of data access schemes referred to as static and dynamic. According to the first (second) scheme a transaction reads all the required data items at the beginning of its processing (on demand during its processing), respectively. After completing its processing, each transaction is checked as to whether it has encountered a data conflict. Validated transactions commit; otherwise, they are restarted. A variant of the regular (silent) commit scheme where a committing transaction notifies conflicted transactions to restart immediately (broadcast commit scheme) is also considered. We use an iterative method to analyze the performance of OCC schemes in the framework of a system with a fixed number of transactions in multiple classes with given probabilities for their occurrence. The iterative method is validated against simulation and shown to be highly accurate even for high data contention. We present graphs/tables, which are used to determine how system performance is affected by: (i) various OCC schemes, (ii) transaction size, i.e., number of data items accessed, (iii) number of transactions, (iv) the distribution of transaction processing time requirements, (v) the throughput characteristic of the system, and (vi) granule placement.

Observations on optimistic concurrency control schemes

Optimistic concurrency control schemes allow uncontrolled access to shared data objects during transaction processing under the explicit assumption that read and write conflicts among transactions are rare events. Before a transaction commits, the DBMS has to validate that no conflict has occurred. Conflict resolution mainly relies on transaction abort. Two different optimistic concurrency control schemes are introduced and compared to each other. The problems of implementing such schemes and their implications on DBMS processing is investigated in some detail. A number of general properties of optimistic concurrency control schemes is derived, and their advantages and drawbacks w.r.t. two-phase locking approaches are discussed.