JDBC-ODBC Bridge Replacement Yields 310% Performance Boost
See how DataDirect's adapter matches up against Sun's offering.
Join the DZone community and get the full member experience.Join For Free
Performance tests reveal that the Progress DataDirect Sequelink JDBC-ODBC adapter is up to 310% faster than the Sun JDBC-ODBC bridge.
The DataDirect Sequelink JDBC-ODBC adapter serves as the trusted replacement for the JDBC-ODBC bridge. Even Oracle Data Integration recommends DataDirect SequeLink, after deprecating the JDBC-ODBC bridge in JDK 8. This adapter is extremely useful when you need connectivity to legacy data sources or when organizations need JDBC connectivity and only ODBC drivers are available.
Conversely, the DataDirect SequeLink ODBC-JDBC adapter serves as the adapter for scenarios where ODBC access is required, and only JDBC drivers exist, such as Apache Solr, Apache Phoenix or Neo4j.
As organizations turn to DataDirect for a JDBC-ODBC bridge replacement, we wanted to provide some performance indications so you can consider a replacement before moving to JDK 8. Our developer evangelist, Saikrishna Teja Bobba, ran the following tests.
Environment and Tools Used
To evaluate the performance of these drivers, I have used the following setup.
- Server: MemSQL 5.1.0 hosted on Cloud VM with 7GB RAM and 4 cores CPU
- Client: Windows 7 64 bit VM with 4GB RAM and 2 Cores CPU
- ODBC Driver: Native MySQL 32 bit ODBC Driver 5.3.6
- Measuring Tool: Apache Jmeter 3.0
Plot of Time taken (in milliseconds) vs. Number of records that are pulled through using JDBC-ODBC bridge drivers. (Lower is Better.)
From these numbers, you can see that the DataDirect Sequelink JDBC-ODBC adapter is three times faster than the Sun JDBC-ODBC drivers. Along with Oracle recommending us as a replacement to the Sun JDBC-ODBC bridge, we add production ready quality and improved performance for your application as part of this transition.
We tested these JDBC-ODBC bridges with data pull loads of 300,000, 600,000, and 1 million records. For each load, the test was performed three times and an average value was calculated. This testing mechanism was configured in Apache Jmeter and the average latency values were recorded from there.
Why Did We Use a Native MySQL ODBC Driver, Instead of Our In-House DataDirect MySQL ODBC Drivers?
The reason we used the Native MySQL ODBC driver was to have a fair competition between the drivers. We didn't want people to think that our Sequelink JDBC-ODBC bridge driver had an advantage because we used our in-house MySQL drivers.
If you're wondering whether our DataDirect MySQL drivers are in fact faster than Native MySQL drivers — the answer is yes. You can read this blog where we published the results of a comparison between the DataDirect MySQL driver and the Native MySQL driver.
Why Did We Use MemSQL as a Data Source, When It Already Has a JDBC Driver?
The main purpose of this performance test was to calculate how fast these JDBC-ODBC bridges could communicate JDBC calls to the ODBC API and get the data. To put it simply, we are only interested in calculating the overhead when using these bridges. We are not interested in how fast these data sources can serve the results.
Learn more and start a trial of the DataDirect Sequelink JDBC-ODBC adapter.
Published at DZone with permission of , DZone MVB. See the original article here.
Opinions expressed by DZone contributors are their own.