Oracle is Killing MySQL; Winter Follows Autumn

Although Sun Microsystems was a victim of its own chronic mismanagement, I think that few people must have looked upon its takeover by Oracle as anything other than a disaster for everything Sun had achieved in its 28-year history.

Company mergers are often referred to in the iconoclastic online journals as “Borging‚” referring to the enemy of the Star Trek: The Next Generation. For the 1 of you reading this who is not familiar with the Borg, they take over (“assimilate”) other races, making them part of the collective, and in the process eliminating anything that had previously differentiated the Borged race. Unfortunately, in the case of Oracle’s takeover of Sun, the analogy appears utterly accurate.

Almost certainly a contributing factor in their relative corporate success levels, Sun had perpetrated, or adopted, a number of technologies for the benefit of the community. Java is one, MySQL another. Whereas, on the other side of the takeover fence, Oracle has always been all about the money.

To say that the open-source MySQL project has completely revolutionised the internet would not be mere hyperbole. Millions of web sites, from the mighty Facebook to my own paltry corner, run on MySQL servers, and therein lies the very attraction of the product. The fact is that MySQL is almost limitlessly scalable. From a single database mouldering away on a 10-year-old PC in a garage up to the powerhouse for a multinational website with 500 million users, MySQL can grow with your organisation. In the process, it passes through the comfort zones of other database products, ones which may be more familiar and/or comfortable for corporate users. Microsoft’s SQL server is pretty well unbeatable for driving a 2,000-user corporate application, and Oracle’s own products specialise in huge corporate data warehouses.

Part of this specialisation comes from the native capability of the products. Part of it comes from the platforms on which the products will run. Part of it comes from the complexity of architecting a database solution that works within the cost constraints. And part of it from the purchase cost of the product itself.

And this is where Oracle’s money-grabbing ways threaten the database ecosystem.

Microsoft’s SQL Server is a good midrange tool. It runs on Windows servers, on Intel and AMD hardware, and supports clustering for high availability. It is a strong product for taking an organisation from the realisation they need a database through to running significant-sized corporate applications. But it can be complex to buy, and if you get the buying choices wrong, you’re in for a load of cost that you don’t need.

To offset some of this, Microsoft offers SQL Express, which is free, but limited. It’s a helping hand onto the SQL Server ladder.

Oracle’s database product set has a number of advantages over SQL Server in that it will run on almost anything, and its track record in massive data warehouses is more compelling than that of SQL Server. It is, however, both complex and expensive to buy. Each component is costly, and there are lots of them.

Where does Sun come into this? Well, Sun’s nurturing of the open source MySQL platform had given MySQL corporate credibility. It had become possible to obtain the open source product, and then to purchase support for it at a reasonable price. The sort of price that a corporate customer without much of a budget could afford. With a ready supply of self-taught LAMP (Linux, Apache, MySQL, PHP) developers out there in internet-land, corporates could get up and running for a modest investment and then grow their capability in line with their requirements. Are growing MySQL implementations a significant threat to new Oracle installations? Or are they an opportunity for Oracle to sell in right-sized support solutions that give customers a feeling of value for money?

When Oracle borged Sun, the optimists thought that MySQL could become Oracle’s SQL Express: the free offering to get customers onto the ladder (or “tied in”if you’re not feeling charitable!).

The optimists were wrong though. Oracle remains all about the money coming in. It has canned low-cost support options, and pulled components such as Enterprise Monitor from the low-end commercial editions of the product. Add-on technologies, such as MySQL Enterprise Backup will not be open-sourced, and have to be purchased.

It seems that Oracle views MySQL implementations as a threat to sales of its enterprise-level database products, and does not value SME revenue streams. It would appear that Oracle is happier to drive customers to Microsoft’s waiting arms than to interact with those customers with anything other than its flagship product.

It’s an approach that is remarkably short-sighted. By applying pressure to the community Oracle is causing key personnel to depart and form new alliances. Already a new start-up exists “SkySQL” that has many of the original guiding lights from the MySQL project. Some may argue that this is part of a natural process, that Oracle would be commercially naïve to retain a non-profit organisation under its wing that actively causes customers to not buy its commercial product.

But with Oracle running down Java, commercially squeezing the open source database, and failing to turn around the server hardware business, the question that remains unanswered is that of why Oracle took over Sun in the first place.

Was it ego inflation through corporate carrion consumption for Larry Ellison?