Udai Shanker

Distributed real time database systems: background and literature review

Today’s real-time systems (RTS) are characterized by managing large volumes of dispersed data making real-time distributed data processing a reality. Large business houses need to do distributed processing for many reasons, and they often must do it in order to stay competitive. So, efficient database management algorithms and protocols for accessing and manipulating data are required to satisfy timing constraints of supported applications. Therefore, new research in distributed real-time database systems (DRTDBS) is needed to investigate possible ways of applying database systems technology to real-time systems. This paper first discusses the performance issues that are important to DRTDBS, and then surveys the research that has been done so far on the issues like priority assignment policy, commit protocols and optimizing the use of memory in non-replicated/replicated environment pertaining to distributed real time transaction processing. In fact, this study provides a foundation for addressing performance issues important for the management of very large real time data and pointer to other publications in journals and conference proceedings for further investigation of unanswered research questions.

SWIFT--A new real time commit protocol

Although there are several factors contributing to the difficulty in meeting distributed real time transaction deadlines, data conflicts among transactions, especially in commitment phase, are the prime factor resulting in system performance degradation. Therefore, design of an efficient commit protocol is of great significance for distributed real time database systems (DRTDBS). Most of the existing commit protocols try to improve system performance by allowing a committing cohort to lend its data to an executing cohort, thus reducing data inaccessibility. These protocols block the borrower when it tries to send WORKDONE/PREPARED message [1, 6, 8, 9], thus increasing the transactions commit time. This paper first analyzes all kind of dependencies that may arise due to data access conflicts among executing-committing transactions when a committing cohort is allowed to lend its data to an executing cohort. It then proposes a static two-phase locking and high priority based, write-update type, ideal for fast and timeliness commit protocol i.e. SWIFT. In SWIFT, the execution phase of a cohort is divided into two parts, locking phase and processing phase and then, in place of WORKDONE message, WORKSTARTED message is sent just before the start of processing phase of the cohort. Further, the borrower is allowed to send WORKSTARTED message, if it is only commit dependent on other cohorts instead of being blocked as opposed to [1, 6, 8, 9]. This reduces the time needed for commit processing and is free from cascaded aborts. To ensure non-violation of ACID properties, checking of completion of processing and the removal of dependency of cohort are required before sending the YES-VOTE message. Simulation results show that SWIFT improves the system performance in comparison to earlier protocol. The performance of SWIFT is also analyzed for partial read-only optimization, which minimizes intersite message traffic, execute-commit conflicts and log writes consequently resulting in a better response time. The impact of permitting the cohorts of the same transaction to communicate with each other [5] on SWIFT has also been analyzed.

Dependency Sensitive Shadow SWIFT

This paper proposes a new commit protocol for real time distributed database systems, dependency sensitive shadow SWIFT (DSS-SWIFT) protocol where the cohort forks off a replica of itself called a shadow, whenever it borrows dirty value of a data item, and if, the created dependency is abort type as compared to creating shadow in all cases of dependency in shadow PROMPT. Also the health factor of cohort is used for permitting to use dirty value of lender rather than health factor of transaction as whole. The performance of DSS-SWIFT is compared with PROMPT, 2SC and SWIFT protocols for both main memory resident and disk resident databases with and without communication delay. Simulation results show that the proposed protocol improves the system performance up to 7% as transaction miss percentage

A memory efficient fast distributed real time commit protocol

Most of the past researches [1], [2], [3] investigate the behavior of distributed real time commit protocols either under update or blind write model. The effect of both types of models has not been investigated collectively. These protocols also require a considerable amount of memory for maintaining temporary objects (data structure) created during execution of transactions and block the WORKDONE message if cohort is dependent. This paper presents an optimized distributed real time commit protocol (MEFCP) based on new locking scheme and write operation divided into update and blind write. The proposed protocol optimizes the memory required for maintaining the transient information of lender & borrower [1]. It also sends the WORKDONE message if borrower has locked the data in mode 2 only. We also compared MEFCP with PROMPT and 2SC commit protocols through simulation.

SPMC-PRRP: A Predicted Region Based Cache Replacement Policy

Earlier cache replacement policies are solely based on the temporal features of client’s access pattern. As mobile client moves to different locations, their access pattern shows not only temporal locality rather it shows spatial locality as well. The previous policies have not evolved any accurate next location prediction policy that can be used in cost computation of data items. To overcome this limitation of previous policies, this paper proposes a Cache Replacement Policy (SPMC-PRRP) based on mobility prediction. To predict accurate future location of moving client, mobility rules have also been framed based on similarities between user’s movement data. Here, removal of noise (random movements) in mobile users’ profiles in trajectories has been done to be used in above policy. Simulation results show that proposed policy achieves up to 5% performance improvement compared to previous well-known replacement policies via improving the accuracy in next location prediction.

On Using Priority Inheritance-Based Distributed Static Two-Phase Locking Protocol

Two-phase locking with high priority (2PL-HP), a well-suited concurrency control protocol for distributed real-time database systems (DRTDBS) because of being free from priority inversion problem, is used for accessing data items to resolve conflicts among the concurrently executing transactions. However, it suffers from the problems of wastage of system resources responsible for degrading the system performance. In DRTDBS, our basic aim is to minimize the number of transactions missing their deadline. In this paper, static two-phase locking with priority inheritance (S2PL-PI) protocol has been proposed specifically to minimize the wasted system resources, i.e., CPU and data items by avoiding unnecessary abort of transactions by optimal use of priority inheritance mechanism. A DRTDBS is simulated for comparison of the performance of S2PL-PI protocol with previous other protocols, and results confirm the significant improvement in system performance.

CART: A Real-Time Concurrency Control Protocol

The Two Phase Locking with High Priority (2PL-HP) concurrency control protocol addresses the transaction scheduling issue in a distributed real-time database system (DRTDBS). Although the 2PL-HP protocol is free from priority inversion, it may suffer from the problems such as deadlock, cyclic restart, and starvation of lengthy transactions. In this paper, a Controlled Avoidance of deadlock and starvation causing Resourceful Conflict resolution between Transactions (CART) concurrency control protocol has been proposed to minimize the transactions miss percentage by reducing the wastage of system resources through avoiding the deadlock due to controlled locking and starvation to some extent by ensuring a fairness in the allocation of resources for their completion. DRTDBS is simulated and CART outperforms as compared with previous other protocols.

CELPB: A Cache Invalidation Policy for Location Dependent Data in Mobile Environment

Location dependent information services (LDIS) can be characterized as the applications that coordinate a cell phones area or position with other data to give enhanced value of services to the client at right place in the right time from anywhere. In this paper, an algorithm Caching Efficiency with Next Location Prediction Based (CELPB) has been developed that uses a newly developed metric i.e. caching efficiency with next location prediction (CELP) for the computation of valid scope in prediction interval. This metric takes account the future movement behavior of client with the help of Sequential Pattern Mining and Clustering. The mobility rules have also been framed for the prediction of an accurate next location, which can be used in estimating the future movement path (edges) of client if he reached in valid scope area of any data item. Simulation results show that proposed policy achieves up to 10 percent performance improvement compared to earlier cache invalidation policy (CEBAB) for LDIS.