Main synopsis
Download as DOCX, PDF1 like510 views
This document provides a synopsis for a minor project comparing the efficiency and performance of different concurrency control algorithms in databases. The project was submitted by four students to their professor. The objective was to analyze research papers on concurrency control algorithms and identify ways to improve performance and resource utilization. The introduction discusses four locking algorithms, two timestamp algorithms, and one optimistic algorithm studied, and concludes that locking algorithms provide the best performance. Technical details describe a queuing model used to simulate the performance of transactions moving through startup, concurrency control, object access, and blocked queues.
Main synopsis
- 1. Minor Project Synopsis Page 1 Minor Project Synopsis Efficiency and performance of “Concurrency Control Algorithm in Databases” Submitted to:Ritu Sharma Submitted by: Anurag Sharma, 70 Narender Sharma, 69 Sudhanshu Gupta, 61 Vishal Singh, 68
- 2. Minor Project Synopsis Page 2 Objective: This is a research oriented work which has led to the development of different version of concurrency control algorithm, most of which are based on one of the three mechanisms: Locking, Timestamps, and commit-time validation (Called optimistic algorithm) Aim: Our Team traversed through some couple of research papers and analyzed that concurrency control algorithm which is used in several databases (like oracle, Mysql, Sql Server, PostGre Sql etc.) can be improved by changing certain fragments to provide better performance and proper resource utilization.
- 3. Minor Project Synopsis Page 3 Introduction: The basic concurrency control algorithm studied includes four locking algorithm, two timestamp algorithm, and one optimistic algorithm. They conclude that locking algorithm provide the best performance. Technical Details: Queuing Model: Central to the detailed simulation approach used here is the closed queuing model of a single-site database system; this model is an extended version of the model of Ries. There are a fixed number of terminals from which transactions originate. When a new transaction begins running, it enters the Startup queue, where processing tasks such as query analysis, authentication, and other preliminary processing steps are performed. Once this phase of transaction processing is complete,
- 4. Minor Project Synopsis Page 4 the transaction enters the concurrency control queue (or cc queue) and makes the first of its concurrency control requests. If this request is granted, the transaction proceeds to the object queue and accesses its first object. if more than one object is to be accessed prior to the next concurrency control request, the transaction will cycle through this queue several times. When the next concurrency control request is required, the transaction re-enters the concurrency control queue and makes the request. It is assumed for convenience that transactions which read and write objects perform all of their reads before performing any writes. If the result of a concurrency control request is that the transaction must block, it enters the blocked queue until it is once again able to proceed. If a request leads to a decision to restart the transaction, it goes to the back of the
- 5. Minor Project Synopsis Page 5 concurrency control queue after a randomly determined restart delay period of mean restart_delay;it then begins making all of its concurrency control requests and objects accesses over again.