Select industry sectors

Sector IQ

The wider implications of Oracle's impending ownership of MySQL

Analyst: Matt Aslett
Date: 12 May 2009
Email This Report: to Colleagues »» / to yourself »»
451 Report Folder: File report »» / View my folder »»

Just under a year ago, we speculated about how the database market would respond to Sun Microsystems' (Nasdaq: JAVA) acquisition of open source database vendor MySQL for $1bn. Twelve months later, ownership of MySQL is about to change hands once again, with Sun having agreed to be purchased following a $7.4bn bid from database, middleware and applications giant Oracle (Nasdaq: ORCL). We have already provided our first impressions of the pending transaction, but Oracle's strength in the enterprise database market, combined with MySQL's rapid rise in the Web database space, makes a closer look at the implications for the data management sector worthwhile.

There is the question of just what Oracle will do with MySQL, the open source database that has carved a substantial niche in the database market by targeting Web applications, rather than the enterprise space dominated by Oracle, IBM (NYSE: IBM) and Microsoft (Nasdaq: MSFT). Will Oracle pay MySQL the ultimate compliment by admitting that its competitive threat is its database business and killing it off, or will it use MySQL as a weapon to target its rivals? Then there is the matter of the ecosystem of vendors that has developed around MySQL, offering additional functionality to make the database more suitable for use in transactional and data-warehousing environments – which is Oracle Database territory. Will Oracle look favorably on the MySQL partners, or take the opportunity to undermine potential competitors?

What will Oracle do with MySQL?

It's easy to assume that Oracle will see its acquisition of Sun as an opportunity to attempt to kill off MySQL or, at the very least, let it wither on the vine. There are a number of reasons why we believe MySQL is more valuable to Oracle alive than dead. The first of these is that MySQL is largely complementary to Oracle's Database products, rather than directly competitive. MySQL was created as a lightweight database for Web and departmental applications that do not require the functionality offered by the likes of Oracle and IBM's DB2, and the company grew rapidly by avoiding direct competition with those vendors.

One of the reasons why Sun saw value in MySQL beyond its direct revenue was its community of users, an estimated 11 million installations and some 10,000 paying customers worldwide at last count, and the ability to sell other products and services into those accounts. We believe Oracle will also see value in the opportunity to gain entry into accounts that have previously avoided its database offerings and sell other products and services.

Over the years, more enterprise features have been added to MySQL, and there is now some overlap, but the adoption of MySQL as a lightweight, low-cost, scale-out database for the Web tier means that it is focused on environments and workloads that are, for the most part, unsuitable for Oracle's scale-up approach and bells-and-whistles enterprise functionality. If Oracle were to deliberately let MySQL die, its users would be unlikely to go running into the arms of the Oracle Database sales team but would instead look for open source alternatives like PostgreSQL or Ingres, or Microsoft's SQL Server. MySQL competes much more directly with SQL Server at the Web and departmental tier, and we would expect to see Oracle playing to MySQL's strengths and using it as a competitive weapon to target its rivals in those markets.

Then there is the fact that, because MySQL is an open source project, Oracle could not truly kill it off even if it wanted to. There are already a number of alternative development branches, or forks, of MySQL, and were Oracle to antagonize the MySQL user base, we would expect one of those to emerge to take MySQL's place (MariaDB, the version created by original MySQL creator Monty Widenius, would likely be a frontrunner). Oracle has little to gain by driving MySQL users into the arms of its rivals or 'MySQL AB version 2.0,' and the company's executives should be well aware of this.

Oracle's database acquisition track record

It's also worth considering Oracle's track record in terms of buying open source database assets, and its attempts to acquire MySQL. Oracle had made at least one effort to purchase MySQL before Sun snapped it up. Was its plan to simply take out a rival? Perhaps. But we previously noted that MySQL's Web-application capabilities for distributed architectures would also have sat nicely alongside Oracle's transactional-application capabilities for clustered or stand-alone big-box server deployments.

Many people assumed that disruption of MySQL was Oracle's plan when, in October 2005, it acquired Innobase, a Finnish company responsible for developing the open source InnoDB storage engine that was – and remains – the default database engine for MySQL, providing it with ACID-compliant transaction support, among other things.

To the surprise of many, however, the InnoDB team continued its development work within Oracle and in April 2006, the two companies extended MySQL's license to support and sell the InnoDB storage engine. Likewise Berkeley DB, the embedded open source database, was acquired by Oracle in February 2006 along with Sleepycat Software. While it perhaps does not enjoy the profile it had prior to the pickup, Berkeley DB continues to be developed and remains Oracle's primary offering in the embedded database space. Oracle's ownership of MySQL would unite the InnoDB and MySQL development teams for the first time and, arguably, enable the company to bring an even stronger product to market, especially in arenas where the Oracle Database does not necessarily do as well.

Oracle has made relatively few purchases related to its core database business in recent years, preferring instead to use money to fund consolidation in the middleware and application sectors. Its most recent data management acquisition was of data-caching and distributed data management vendor Tangosol in March 2007. Tangosol's products continue to be developed and sold under the Coherence brand, and in fact all of the data management offerings acquired by Oracle in recent years continue to be developed and marketed as stand-alone products.

Oracle's recent database-related acquisitions

Source: The 451 M&A KnowledgeBase *451 Group estimate

Of course, adding MySQL to the Oracle portfolio will be easier said than done, and the acquirer will have to focus its attention very carefully on the sales process. In order for the strategy we suggest above to work, the vendor would need to have separate Oracle Database and MySQL teams, and managing potentially competing sales teams is never easy. Oracle would have to ensure that turf wars do not develop. There is also the matter of the exodus of MySQL senior executives, including former CEO Marten Mickos and chief technical officer Monty Widenius. The company would have to retain as many of the remaining MySQL executives and developers as possible to avoid MySQL being diluted into the database business, which will naturally be more focused on the Oracle Database.

Wider implications

Fears about the long-term future of InnoDB following its acquisition by Oracle prompted MySQL to accelerate its storage engine partner program to encourage third parties to create storage engines that would boost the performance and suitability of MySQL for different application workloads. The strategy has been successful, with a number of vendors seeing an opportunity to build a business model around serving the MySQL user base with additional functionality. Examples include Kickfire and Infobright in data warehousing, ScaleDB in shared-disk clustering, Tokutek in Web-application querying, and Schooner Information Technology and Virident Systems in caching appliances.

While Oracle might be happy with MySQL complementing its database business, it is unlikely to be as impressed with third-party vendors increasing MySQL's competitiveness against the Oracle Database, and we believe Oracle will focus all of its attention on InnoDB and MySQL's in-house storage engines rather than on providing a business opportunity to potential competitors. For the commercial arrangements between these vendors and Oracle to survive, they will have to show that they can provide value to MySQL without impacting Oracle. For some that will be easier said than done.

Sun also previously offered support for PostgreSQL, an open source database that is more directly competitive with Oracle and IBM. We would expect the remains of that business to be discontinued, along with Sun's relationship with Greenplum, the data-warehousing vendor that is itself built around PostgreSQL. Sun was a close ally for Greenplum, providing the hardware for its data-warehouse appliances, and invested in the company. Given Oracle's position in the data-warehousing market, we would expect this relationship to be extinguished. Infobright and Kickfire might be more lucky since they are specifically targeted at the MySQL installed base, but we have our doubts about whether Oracle will look kindly on them.

The hardware angle

The data-warehousing space also provides an interesting perspective on Oracle's potential plans for Sun's hardware business. The Hewlett-Packard (NYSE: HPQ) Oracle Database Machine and the HP Oracle Exadata Storage Server, introduced in late 2008, was an admission by Oracle that sometimes the answer to a software problem (in this case data bandwidth and query performance) is sometimes hardware. Exadata is a storage software and hardware product designed to run the Oracle query engine and improve the overall performance of data-warehousing deployments by taking query processing away from the database engine. There are questions about whether Oracle wants to become a full-fledged systems company and commit to long-term hardware and processor research and development. A compromise could see Oracle selling off the assets that it does not see value in and retaining, or at the very least licensing back at a very competitive rate, the assets that it could use to improve the performance of its various software offerings.

The impact on further M&A

The dominance of Oracle, IBM and Microsoft in the database market means that there have been relatively few database acquisitions in recent years, and that there are relatively few choices for Oracle's rivals to respond to its ownership of MySQL. We reported earlier that MySQL was very much the crown jewel of the open source database world because its focus on Web-based applications, its lightweight architecture and fast-read capabilities, and its brand made it potentially complementary technology for all of the established database vendors.

The same cannot be said of Ingres or EnterpriseDB (based on PostgreSQL), both of which are targeted at traditional enterprise software workloads and are more competitive with the Oracle Database than MySQL. IBM and Microsoft would see little value in either of them, and while the Redmond giant has shown itself to be more accepting of open source in recent years, we believe owning an open source database would be stretching that tolerance too far.

IBM had seemed like a potential buyer for EnterpriseDB since it joined Charles River Ventures, Fidelity Ventures and Valhalla Partners in a $10m series C funding round for EnterpriseDB in March 2008; however, Big Blue recently licensed EnterpriseDB's Oracle-compatibility capabilities for inclusion in DB2, which suggests that its interest was in the company's proprietary extensions rather than the underlying PostgreSQL technology. Besides, IBM already has multiple databases in its portfolio thanks to its purchases of Informix and Solid Information Technology. PostgreSQL would add little of additional value.

A more likely suitor for database technology would be Linux and open source middleware provider Red Hat (NYSE: RHT). We have repeatedly stated our view that the database is a gaping hole in Red Hat's software portfolio – just as the company's executives have repeatedly maintained that they do not see the need to own a database. We think this is due, in part, to Red Hat not wanting to upset its partnership with Oracle, which remains strong even after Oracle launched an alternative Linux support offering with Unbreakable Linux and repackaged Red Hat Enterprise Linux as Oracle Enterprise Linux (OEL). However, we believe the situation would change if Oracle were to use its ownership of MySQL in an attempt to drive users away from Red Hat toward either OEL or perhaps Sun's Solaris Unix OS. Either Ingres or EnterpriseDB would be potential targets for Red Hat, although EnterpriseDB's proprietary extensions do not fit Red Hat's 100% open source business model and would need to be relicensed.

Beyond MySQL, if Oracle were to retain Sun's hardware assets, that might encourage hardware vendors like EMC (NYSE: EMC) and HP to consider acquiring a database. We stand by our previous assessment that EMC missed out on a great opportunity to add a much-needed relational database to its arsenal by not landing MySQL and that a database would be a welcome addition to HP's portfolio. Ingres and Sybase (NYSE: SY) would be of interest to both given their strong enterprise credentials. Ingres would be considerably cheaper and would need bolstering with a significant sales team to offer the sort of competition that HP and EMC would be looking for, while Sybase is the finished package and offers more in terms of its mobility assets, but would come at a price.