MySQL Moves On ClusteringMySQL Moves On Clustering
MySQL Cluster combines the MySQL database with a distributed clustering technology to insure that data is always available.
April 14, 2004
With its aim on the enterprise market, MySQL AB on Wednesday debuted a preview edition of a clustered version of its open-source database, and said it would push out a production version in the third quarter of the year.
The goal, said the Swedish-based open-source developer, is to expand its market to customers that require a high-availability database, but don't want to shell out the money for higher-priced software from database providers such as Oracle, IBM, and Microsoft. "The companies that are using MySQL came back to us, and told us that they were starting to run high-volume sites that simply can't go down," said Edwin DeSouza, the director of product management at MySQL "They needed the high availability delivered by clustering." Rather than competing directly with Oracle, IBM, and Microsoft, DeSouza said, MySQL's intent is instead to stick with what it's known for -- reliable, low-cost databases -- but to also "grow with our customers." Dubbed MySQL Cluster, the offering combines the MySQL database with a distributed clustering technology to insure that data is always available, a key to companies running any database for mission-critical applications, such as handling online transactions. Like databases from IBM and Microsoft, it operates in memory, which means that an enterprise doesn't have to purchase new hardware, such as an expensive storage area network. The clustering technology, as well as its engineering team, was acquired from Swedish telecomm manufacturing giant Ericsson, said Vinay Joosery, the product manager for MySQL Cluster. "We took a proven clustering solution from Ericsson, one that's widely used in the telecom space, and made it work with MySQL," said Joosery. MySQL Cluster claims 99.999 percent availability through placing databases on multiple nodes. If one fails, another can come online in under a second, Joosery said from the company's user conference in Orlando. While MySQL Cluster can operate on as many as 64 nodes -- and has been tested and benchmarked on 48 -- Joosery said the most likely applications would be on far fewer systems. "If you want to cluster with off the shelf components, Linux boxes and Ethernet or Gigabit Ethernet, you can really only scale up to 16 nodes," he said. "But we expect that most sites will deploy two to four nodes." Swedish Internet service provider Bredbandsbaloget, for instance, which has been running MySQL Cluster to guarantee broadband access and voice-over-IP (VoIP) services to its customers, currently uses it on a four-node cluster. While DeSouza maintained that MySQL Cluster won't compete with Oracle, IBM, and Microsoft in the Fortune 100 or Fortune 500 space -- "We build databases for the masses," he said repeatedly, "and this is clustering for the masses" -- he didn't hesitate to point out the lower cost of his clustering solution. "We'll be orders of magnitude cheaper," he said. A preview of MySQL Cluster can be downloaded now under the open-source GPL license from the MySQL Web site, but when it goes final in the third quarter, both a GPL version and a commercially licensed version will be available. "The price of the commercial license will be significantly under $5,000 per CPU," DeSouza promised. He compared that with Oracle, Microsoft, and IBM, whose prices range from $20,000 to $32,000 per processor for their clustered databases. MySQL Cluster runs across a variety of hardware and operating system platforms, including Linux, Windows, Solaris, and Mac OS X. One analyst saw a definite niche for MySQL Cluster. "The new clustering solution makes MySQL look attractive to customers that are looking for enterprise level DBMS features at low cost," said Noel Yuhanna, a senior analyst at Forrester Research, in a statement accompanying the roll-out.
About the Author(s)
You May Also Like
Perspectives on Security for the Board - 3rd Edition
The Forrester Wave™: Vulnerability Risk Management, Q3 2023
Cloud Security Maturity Model: Vision, Path, Execution
Responsible data use: Navigating privacy in the information lifecycle
Checklist: Top 6 Considerations to Optimize Your Digital Acceleration Security Spend