'We don't see ourselves competing with Oracle and IBM. There's a need for a commodity database that's fast and easy to use. We don't see IBM and Oracle trying to be in that space.'
Mårten Mickos, the CEO of MySQL AB, is steering his privately-held company one step at time. The open source database company has just $5 million in revenues last year, mostly from licensing of a commercial version of the software. But Mickos believes his company could become the next Red Hat…or Oracle. Benchmark Capital invested nearly $20 million in the company in June, and enterprise software giant SAP inked a deal with MySQL that will accelerate the fledgling database company's journey into the enterprise.
Today, MySQL has an estimated 4 million installations, but the company is generating revenue only from less than one percent of those sites. Nonetheless, Mickos believes the "commodity" database market will continue to grow rapidly, and over time the company can add features required for heavy duty enterprise applications.
In our interview, Mickos talked about how MySQL is dealing with copyright and patent infringement issues that the SCO suit has surfaced, the company's business model, and how he thinks about the big three: Oracle, IBM, and Microsoft.
ZDNet: Open source has been under scrutiny, as we have seen with SCO's suit versus IBM and its extension to Linux. How are you protecting MySQL against claims of patent or copyright infringement?
Mickos: First, it's important to know it's completely different to protect against patent and copyright infringement. For copyright, if you follow good business ethics, you will be safe. MySQL is developed almost all in-house, so we can be 100 percent clean on the copyright front.
With patent infringement, we have the same challenge as Microsoft, Oracle, or any other software company. To know you are not infringing on patents, you have to look through tens of thousands of patents. It's commonly known that there are so many software patents that it's practically impossible to have 100 percent assurance for any software vendor. Because of the public scrutiny, open source is probably doing better than closed source vendors.
ZDNet: Do you have any specific procedures in place to evaluate code for possible patent infringements?
Mickos: We don't have a separate division looking at patents, but some of our development guys spend some of their time looking at patents and making sure we aren't using those [infringing] technologies?
ZDNet: How do you discern when you might have code that needs to be vetted against existing patents?
Mickos: It's based on personal judgment. In our company, we are building databases using technology that has been around for 10 or 20 years. We are not out creating completely new functionality. The technology is older and mainstream, so not much is patentable in it. We aren't interested in specialized features that could more easily be infringing.
ZDNet: What are some of specialize features you aren't interested in?
Mickos: We are solving problems with mainstream, not specialized, features. For example, object oriented databases may have some patentable stuff, but we aren't interested.
ZDNet: What about clustering--is that specialized?
Mickos: No. It will be a mainstream thing and the majority of our customers will need it. I wouldn't be surprised if someone filed a patent for some of that technology.
ZDNet: Would you file patents on any of your technology?
Mickos: Our company policy is to be against patents. They are not needed for software business, and so far we have not filed for any patents. Red Hat has filed for defensive patents, and I wouldn't exclude that.
ZDNet: What are the current statistics for the company?
Mickos: From the top level, we supply about 35,000 downloads per day; if we add what others are distributing, it's about 20 million downloads per year. We estimate there are 4 million active installations of MySQL, which would put us on par with Oracle and IBM. We have about 4,000 paying customers, so that puts us at one in one thousand [0.001 percent of the estimated 4 million installations]. We had about $5M in revenue last year, and we are on target for doubling this year.
ZDNet: You recently received a $19.5 million investment from Benchmark Capital Investment and inked a deal with SAP. Are these steps toward making MySQL more legitimate as a company?
Mickos: We have been legitimate for long time but not in the enterprise market. You could compare what IBM did with Linux three years ago to what SAP did with us. It's a big event, but the effects will be seen over a number of years.
ZDNet: What will be the first results from your partnership with SAP?
Mickos: First, we are taking over the SAP DB and making the two products [MySQL and SAP DB] interoperable, and over time we will implement in MySQL all the features from SAP DB to support SAP applications.
ZDNet: Have you had talks with other enterprise applications companies, such as PeopleSoft and Oracle?
ZDNet: You mentioned you had revenues of $5 million last year. How much of it is licensing versus services?
Mickos: We have a simple pricing of $440 per server for commercial license fees. We don't ask how many CPUs or servers. We don't force you to pay a maintenance fee. Less of our revenues are for support, training and consulting. Ideally, we would sell high-margin product--commercial licenses--rather than low margin products like services. Our unit prices are low and should be compared to $50,000 to $100,000 for Oracle or DB2. Ours is really a commodity pricing.
ZDNet: Microsoft is also a doing bottom-up approach in acquiring customers. How does that differ from your strategy?
Mickos: We see Oracle, IBM, and Microsoft all belonging to the same group. They are legacy database vendors. Their products were designed in the 1970's and are 10 or more years older than ours. They didn't think about open source or have the same priorities that we have. They come out of a completely different culture. Our number one priority is performance, and theirs is features.
ZDNet: Don't customers want features in their databases?
Mickos: Yes, and we recommend they go to those other database companies.
ZDNet: You characterize Oracle, IBM, and Microsoft as legacy databases, but how old is SQL?
Mickos: How old is SQL? It was developed by IBM in 1970's.
ZDNet: Then why isn't MySQL legacy like Oracle, Microsoft, and IBM?
Mickos: The legacy databases were optimized for client server use, but today most new applications are architected for the Web. They are intrinsically different, and built for a different environment. They are not easy to set up and get running; it's a matter of days and not minutes as in our case.
ZDNet: How are the other databases not Web savvy? Haven't they caught up?
Mickos: Yes, but they are still legacy. Things like the way connections are established and how much memory it takes. With MySQL, we can have more connections in memory space and create them more easily and have better performance in a Web environment.
ZDNet: Is the primary selling proposition that MySQL is free, or very low cost?
Mickos: It's different for different people. Those who use MySQL use four decision criteria. Price is one, but performance, reliability and convenience also factor into a decision. The fact that the software is free doesn't lessen the value of the other criteria. We don't see ourselves competing with Oracle and IBM. There's a need for a commodity database that's fast and easy to use. We don't see IBM and Oracle trying to be in that space.
ZDNet: You said you are a commodity provider of database software, but you also talked about moving into the enterprise. How do you get there?
Mickos: We are adding new features, just like Linux. The big business opportunity is that the commodity market itself will grow--the legacy market is stagnating and the commodity market is growing.
ZDNet: What do you see as your biggest challenges in the next 12 months?
Mickos: We have to manage our growth--the enemy is just ourselves. We have people on board who have been in growth situations before and know what to do. Statistically, many companies had great starts and got tangled in their own growth. Second, open source is a hot topic and everything looks easy and there are many temptations. We have to stay focused and not let our priorities change.
An interesting challenge is how we can sell to the enterprise. We have millions of users, and we have sneaked into nearly every Fortune 500 company, but we are not necessarily on the CIO's radar. How do you sell enterprises with a unit price of $440? Our sales model can't be based on sending in hordes of salespeople to sell a $50,000 product like Oracle. We have to focus on just sneaking in and being there. In addition, we have to make the buying very simple, with licensing, negotiation, and deployment a no brainer. I think we are up to it, and we will innovate in fine-tuning this model.