Hybrid Concurrency Control Research Articles

A Memory-Disaggregated Radix Tree

Disaggregated memory (DM) is an increasingly prevalent architecture with high resource utilization. It separates computing and memory resources into two pools and interconnects them with fast networks. Existing range indexes on DM are based on B+ trees, which suffer from large inherent read and write amplifications. The read and write amplifications rapidly saturate the network bandwidth, resulting in low request throughput and high access latency of B+ trees on DM. In this paper, we propose that the radix tree is more suitable for DM than the B+ tree due to smaller read and write amplifications. However, constructing a radix tree on DM is challenging due to the costly lock-based concurrency control, the bounded memory-side IOPS, and the complicated computing-side cache validation. To address these challenges, we design SMART , the first radix tree for disaggregated memory with high performance. Specifically, we leverage 1) a hybrid concurrency control scheme including lock-free internal nodes and fine-grained lock-based leaf nodes to reduce lock overhead, 2) a computing-side read-delegation and write-combining technique to break through the IOPS upper bound by reducing redundant I/Os, and 3) a simple yet effective reverse check mechanism for computing-side cache validation. Experimental results show that SMART achieves 6.1 × higher throughput under typical write-intensive workloads and 2.8 × higher throughput under read-only workloads in YCSB benchmarks, compared with state-of-the-art B+ trees on DM.

Hybrid concurrency control protocol for data sharing among heterogeneous blockchains

Hybrid concurrency control protocol for data sharing among heterogeneous blockchains

A Hybrid Concurrency Control Strategy for Mobile Database Systems

The inherent limitations of mobile database systems makes concurrency control an important problem. Several pessimistic and optimistic strategies for achieving concurrency control in mobile database systems are presented in literature. The pessimistic strategies achieve concurrency however they face the problem of blocking, whereas the optimistic strategies may not always keep the data consistent. The conflict resolution strategies may result in starvation of some of the mobile requests. This paper presents a hybrid concurrency control strategy by using the good properties of both pessimistic and optimistic approaches. The results specify that the proposed strategy performs better as compared to both pessimistic and optimistic strategies and reduces the starvation of transaction requests of various mobile devices.

An Adaptive Decision Concurrency Control Algorithm

For the lack of adaptability about the existing concurrency control algorithms, adaptive decision concurrency control algorithm is proposed. ADCC algorithm divides concurrency control process into two phases in: execution authorizing phase and strategy selecting phase. In execution authorizing phase, algorithm compares statistics and effectiveness of transactions to determine the execution order of conflict transactions. In strategy selecting phase, according to transactions’ read/write status and current conflict rate, algorithm selects optimistic/pessimistic conflict resolution strategy adaptively. Such selection mechanism makes ADCC algorithm have high efficiency no matter database system is busy or idle. Simulation experiment proves that ADCC algorithm this paper proposed is superior to classical strict two phases locking algorithm and hybrid concurrency control. So ADCC algorithm performs well in the period of concurrency control.

Supporting Fine-Grained Concurrent Tasks and Personal Workspaces for a Hybrid Concurrency Control Mechanism in a Networked Virtual Environment

A concurrency control mechanism for a networked virtual environment is a key element in many collaborative computer-aided design applications. However, conventional object-based locking mechanisms restrict the behaviors of nonowners, and an attribute-based locking mechanism may produce another problem called task-surprise, which disturbs users' collaboration. In this paper, we propose a hybrid concurrency control mechanism that reduces restrictions of nonowners' behaviors and task-surprises in a networked virtual environment. The proposed method consists of two concurrency control approaches: task-based concurrency control and personal workspaces. The task-based concurrency control approach allows nonowners to do some tasks if they do not conflict with the tasks of the owner of the shared object. The personal workspaces approach provides an independent workspace where a user can manipulate copies of the shared objects. The proposed method was applied to a collaborative level design for a large-scale online game as a case study. We evaluated its performance by experiments and user studies to check acceptance and usability of the proposed method.

An intelligent hybrid optimistic/pessimistic concurrency control algorithm for centralized database systems using modified GSA-optimized ART neural model

Concurrency control is the activity of synchronizing operations issued by concurrent executing transactions on a shared database. The aim of this control is to provide an execution that has the same effect as a serial (non-interleaved) one. The optimistic concurrency control technique allows the transactions to execute without synchronization, relying on commit-time validation to ensure serializability. Effectiveness of the optimistic techniques depends on the conflict rate of transactions. Since different systems have various patterns of conflict and the patterns may also change over time, so applying the optimistic scheme to the entire system results in degradation of performance. In this paper, a novel algorithm is proposed that dynamically selects the optimistic or pessimistic approach based on the value of conflict rate. The proposed algorithm uses an adaptive resonance theory–based neural network in making decision for granting a lock or detection of the winner transaction. In addition, the parameters of this neural network are optimized by a modified gravitational search algorithm. On the other hand, in the real operational environments we know the writeset (WS) and readset (RS) only for a fraction of transactions set before execution. So, the proposed algorithm is designed based on optional knowledge about WS and RS of transactions. Experimental results show that the proposed hybrid concurrency control algorithm results in more than 35 % reduction in the number of aborts in high-transaction rates as compared to strict two-phase locking algorithm that is used in many commercial database systems. This improvement is 13 % as compared to pure-pessimistic approach and is more than 31 % as compared to pure-optimistic approach.

Locking protocols for concurrency control in real-time database systems

Concurrency Control in real-time database systems is complicated by the requirement to maintain database consistency at the same time to minimize the number of transactions missing their deadlines. The scheduling of data in Basic Two Phase Locking (2PL) completely ignores the urgency of a transaction and thus the effectiveness of the adopted real-time resource scheduling protocol is greatly reduced. In Restart based locking protocols (R2PL), same priorities are used for both data and resources scheduling and should have fewer transactions missing their deadlines. However, Restart based protocols sufferred the intrinsic weakness of high restart overhead owing to ensure atomicity of transactions. In this paper, based on their weaknesses, a hybrid concurrency control protocol (H2PL) is proposed. Through performance study, results indicate that it can perform well under different degree of deadline constraint and workload as compared with other real-time locking protocols.

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. >

Hybrid concurrency control for abstract data types

Databases and other transaction-processing systems use concurrency control and recovery algorithms to ensure that transactions are atomic (i.e., serializable and recoverable). We present a new algorithm based on locking that permits more concurrency than existing commutativity-based algorithms. The algorithm exploits type-specific properties of objects; necessary and sufficient constraints on lock conflicts are derived directly from a data type specification. In addition, the algorithm permits operations to be both partial and non-deterministic, and it permits the lock mode for an operation to be determined by its results as well as its name and arguments. We give a complete formal description of the algorithm, encompassing both concurrency control and recovery, and prove that the algorithm satisfies hybrid atomicity, a local atomicity property that combines aspects of static and dynamic atomic algorithms. We also show that the algorithm is optimal in the sense that no hybrid atomic locking scheme can permit more concurrency.

Comparing how atomicity mechanisms support replication

Most pessimistic mechanisms for implementing atomicity in distributed systems fall into three broad categories: two-phase locking schemes, timestamping schemes, and hybrid schemes employing both locking and timestamps. This paper proposes a new criterion for evaluating these mechanisms: the constraints they impose on the availability of replicated data.A replicated data item is a typed object that provides a set of operations to its clients. A quorum for an operation is any set of sites whose co-operation suffices to execute that operation, and a quorum assignment associates a set of quorums with each operation. Constraints on quorum assignment determine the range of availability properties realizable by a replication method.This paper compares the constraints on quorum assignment necessary to maximize concurrency under generalized locking, timestamping, and hybrid concurrency control mechanisms. This comparison shows that hybrid schemes impose weaker constraints on availability than timestamping schemes, and locking schemes impose constraints incomparable to those of the others. Because hybrid schemes permit more concurrency than locking schemes, these results suggest that hybrid schemes are preferable to the others for ensuring atomicity in highly available and highly concurrent distributed systems.This research was sponsored by the Defense Advanced Research Projects Agecy (DOD), ARPA Order No. 3597, monitored by the Air Force Avionics Laboratory Under Contract F33615-81-K-1539.The views and conclusions contained in this document are those of the authords and should not be interpreted as representing the official policies, either expressed or implied, of the Defense Advanced Research Projects Agency or the US Government.