Open Source RDBMS - Seamless, Scalable, Stable and Free

한국어 | Login |Register

Current Events
Join our developers event to win one of the valuable prizes!
posted 4 years ago
viewed 2976 times
Share this article

Better Performance with CUBRID 2008 R2.2

This is a performance review of CUBRID 2008 R2.2, an open source and completely free database management system highly optimized for Web applications. The latest R2.2 version was released at the beginning of May, 2010, just a month ago. Now it is faster than even before.

New Enhancements of CUBRID 2008 R2.2 Release

CUBRID is a comprehensive open-source database management system highly optimized for Web Applications, especially when complex business services process large amount of data and generate huge concurrent requests. By providing unique optimized features, CUBRID enables to process much more parallel requests at much less response time. With CUBRID, companies will benefit from superior performance, stability, scalability, and high availability required by organization to provide non-stop services to their valuable customers.

Besides the fact that CUBRID is much faster than other alternative database systems, especially in the environment of READ intensive transactions, here are a few more reasons why CUBRID is a better option for your business.

Improved INSERT performance

In the latest release of CUBRID 2008 R2.2 the INSERT query performance has been incredibly improved by modifying an algorithm to achieve I/O load balancing when the INSERT operations are concentrated at a certain point of time. The graph below shows a change in Transactions per Second (TPS) over 3 hours in a workload, in which INSERT operations were made on 40 tables. You can see that the TPS is increased by more than 10% and sections with zero TPS are absent as compared to CUBRID 2008 R2.0.

Improved INSERT performance

Increased space reusability

In order to make page spaces to be effectively reused in a workload where the INSERT and DELETE operations are repeated, in CUBRID 2008 R2.2 we have modified the space usage logic. The following table shows the number of pages left unused when executing the INSERT operation after executing the DELETE or DROP operations for a table in which 1.3 million records (500,000 pages) were stored. You can see that the space reusability for CUBRID 2008 R2.2 has been dramatically increased as compared to the earlier version R2.1.

Workload CUBRID 2008 R2.1 CUBRID 2008 R2.2

After DELETE

(pages)

After INSERT

(pages)

Reusability

(%)

After DELETE

(pages)

After INSERT

(pages)

Reusability

(%)

1. INSERT

2. DELETE ALL

3. INSERT
275362 57213 20.8 275362 254850 92.6
1. INSERT

2. DROP

3. CREATE

4. INSERT
275362 50746 18.4 275362 274483 99.7

Enhanced HA feature

Another major performance improvement was reached as a result of Resource Management and Failure Detection features. These are included in the CUBRID HA (High Availability) Feature. Such enhancement allowed us to easier configure HA environment and manage related resources. Before we relied on Linux Heartbeat package to handle the Failure Detection and Resource Management. Now the dependence on Linux Heartbeat package used in the existing version of CUBRID is successfully removed.

For more information on other changes, see What’s New in CUBRID 2008 R2.2.



comments powered by Disqus