Browse Prior Art Database

Method For Centralized Synchronization Agent Transparent Scale Out In Shared Disk Clustered Data Base Management System

IP.com Disclosure Number: IPCOM000198597D
Publication Date: 2010-Aug-10
Document File: 2 page(s) / 24K

Publishing Venue

The IP.com Prior Art Database

Abstract

Some shared disk clustered Data Base Management Systems (DBMS) architectures and implementations utilize a centralized agent in order to synchronize access to various shared data base structures, such as locks and pages. As the DBMS cluster size increases it is possible to exhaust various physical resources that the single centralized synchronization agent requires to process the various synchronization requests, eg. Memory/Processor/Network. The idea presented below is how to avoid the single centralized synchronization agent bottlneck in a transparent fashion.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 53% of the total text.

Page 1 of 2

Method For Centralized Synchronization Agent Transparent Scale Out In Shared Disk Clustered Data Base Management System

Disclosed is a process for transparently scaling out a single centralized synchronization agent in a shared disk cluster database management system (DBMS).

A typical shared disk clustered DBMS architecture and implementation utilizes a centralized agent to synchronize access to various shared data base structures, for example, locks and pages. As the DBMS cluster size increases physical resources the single centralized synchronization agent requires to process synchronization requests, including memory,

rocessor and networ

p

k

resources may be exhausted.

There are typically two options to avoid the single centralized synchronization agent becoming a bottleneck. One option is to scale up physical resources used by the centralized synchronization agent, for example, by adding more processor processing power, memory and network resources. The first option is not always possible for various reasons, such as exceeding the limits of the largest single physical server.

Using another option, scaling out the centralized synchronization agent, by dividing processing across multiple agents on multiple physical resources, for example, using physical servers may be used. The disclosed process is an example of the scale out category.

Current solutions for scale out of the centralized synchronization agent require user intervention to dictate placement of various structures across synchronization agents. The intervention typically requires isolation of objects at a fairly high granularity, such as a table or index level and assignment of all synchronization request processing for the isolated objects to one synchronization agent. The intervention may also require schema changes and data base object reorganization.

A disclosed process for achieving transparent scale out of synchronization agent is provided using the...