Apr 292009

In a strange twist of fate, the annual MySQL Conference was just getting underway in Santa Clara, California, USA, when news hit that Oracle Corporation was acquiring Sun and, along with it, MySQL. That the most aggressive enterprise software company was buying the world’s leading open source database struck like lightning. After all, it had been only about 16 months since Sun had acquired MySQL for US $1 billion, and proponents of the open source database were still griping about Sun’s efforts to steer and develop the technology. So it’s quite understandable that MySQL fans should be a bit apprehensive about what Oracle might have in store for the database. Here’s what I think will happen with MySQL under Oracle’s stewardship.

Almost all of MySQL’s numbers are impressive. It has an installed base of more than 12 million users. Various estimates place MySQL with 49% of the open source database market. MySQL is the leading open source database for online applications and services and is used by virtually every Web 2.0 company, ranging from Google, Yahoo!, Facebook, and YouTube to Flickr, Fotolog, Friendster, Second Life, and Wikipedia. More mainstream enterprise users include Associated Press, Shinsei Bank, and Toys “R” Us. Alcatel-Lucent, Nokia, China Mobile, and Deutsche Telekom use MySQL in telecommunications, and MySQL also sees considerable use with software as a service (SaaS) providers and in the OEM embedded market.

According to Karen Tegan Padir, VP of Sun’s MySQL and Software Infrastructure Group, MySQL gets downloaded more than 70,000 times a day. (In 2006, it received about 50,000 downloads daily.) Approximately 1,000 partners offer products and services around MySQL. They include such open source OS players as Red Hat and Ubuntu, development tools (Eclipse), as well as the leading open source BI, CRM, and office productivity suite vendors, including JasperSoft, Pentaho, SugarCRM, and OpenOffice.

The one key area in which MySQL has problems, however, is getting paying customers. According to former MySQL CEO Marten Mickos, fewer than 1 in 1,000 users ever pay for MySQL. Moreover, according to IDC, MySQL’s sales were approximately US $38 million for 2007. Still, the sheer size of MySQL’s user base alone must look very appealing to Oracle management. It doesn’t take much imagination to predict that Oracle will use MySQL as a way to get a foot in the door with these companies, and then try to sway them over to (paying) versions of Oracle DB.

There are other possible opportunities here. In addition to enterprise applications, there is the cloud. MySQL could serve as an onramp for Oracle as it puts its might and weight behind cloud platforms and services.

Oracle will have to continue to support and enhance the community version of MySQL, including promoting it for Web applications (e-commerce, Web services) where it is most popular now. And, if history offers any clues, there are some good signs here for open source proponents. About three years ago, Oracle acquired InnoDB — the storage engine used by MySQL — and has continued to develop this technology. Much the same can be said for BerkeleyDB, another open source database that Oracle bought in 2006. I’ve not heard any major complaints about Oracle’s stewardship of these tools (if you have, let me know).

Should MySQL fans grow too unhappy with Oracle’s efforts at shepherding the open source database, they could fork it off into another distribution and forgo having to deal with Oracle.

Sun’s goal in acquiring MySQL was to turn it into an “enterprise” database for large companies and government organizations to deploy in more mainstream production applications (i.e., not e-commerce), and then leverage it to sell other Sun products (hardware, middleware, OSs, networking, etc.) and services.

I see it as unlikely that Oracle will continue Sun’s effort to develop and promote MySQL as an enterprise-capable database for one simple reason: Why would Oracle want to offer an open source database that competes with its bread-and-butter, Oracle DB? This is sure to annoy MySQL fans; however, Oracle is likely to do this surreptitiously — for example, by carefully limiting or drawing out the release of enhancements pertaining to scalability, clustering, and so on. I wouldn’t be surprised either to see Oracle offer more commercial licensing for an enterprise-class version of MySQL. Although this really riled open source database proponents last year when Sun tried it (and then backed down), Oracle is a much more cunning company than Sun — especially when it comes to marketing. After all, Oracle didn’t get to be the giant it is now just by being nice.

Some folks have said that they see the possibility of porting Oracle’s advanced database technology into MySQL. Although it’s possible, I see this as doubtful for the same reasons just given.

What about Oracle pushing MySQL for data warehousing? Now this is an interesting question, for I’ve heard estimates that as many as one-third of the MySQL installations in production are for small and midsize data warehouses and data marts (i.e., about 10 TBs or less). In addition, our research indicates that approximately 15% of the end-user organizations doing data warehousing today are using open source databases as their data warehouse database, with MySQL overwhelmingly the most popular open source database used for data warehousing and BI. So, mightn’t Oracle decide to offer appliances for MySQL warehouses? After all, it will have Sun’s hardware to load it on, offering a one-shop, “true” data warehousing appliance packaging hardware, OS, database, query, reporting, OLAP, advanced analytics, and so on.

I definitely see Oracle offering data warehousing appliances running on Sun platforms. But I think it very much more likely that it will feature the Oracle database preloaded and designed for various platform/application scenarios, as this would go more hand-in-hand with Oracle’s overall sales and marketing scheme (i.e., “Oracle Warehouse”). I do, however, see Oracle going after end-user organizations already running existing MySQL-based data warehouses for all the reasons I noted in last week’s Advisor that Kickfire would appeal to such companies. That is because many of the organizations that have built data warehouses running on MySQL have “hit the wall” with these applications (in terms of performance, scalability, data management, constant tuning, etc.) as they are now experiencing increasing data volumes and growing numbers of end users. Thus, expect to see Oracle try to “transition” these users over to the Oracle Warehouse.

Oracle’s buying Sun represents the most significant shift in Oracle’s history because it dramatically extends the company’s catalog of offerings, which now include everything from databases, middleware, and enterprise applications to hardware and OS platforms as well as some of the world’s leading open source software — including Java and MySQL.

I believe that the Oracle leadership fully realizes that it has acquired a prize with MySQL. And the company is going to be very careful to nurture and develop both the technology and the MySQL community, making sure the database continues to be the favorite among developers and e-commerce companies, thus forming strong relationships it can build on to sell other products and services from its very extensive lineup.


 Leave a Reply

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>