Login  |  Register  |  Contact

Solaris

Monday, May 04, 2009

Comments on Oracle’s Acquisition of Sun

By Adrian Lane

On Monday at the RSA conference I learned that Oracle is purchasing Sun Microsystems. I was so busy/exhausted from the conference that I forgot about it until this week. This is pretty exciting! Whether it’s really a good or a bad thing depends upon your perspective. Technology-wise it’s a good match, but the corporate cultures are very dissimilar. I have spoken with a few current Sun employees who are really worried about what life will be like at the Big-O. However I heard very much the same concern from many PeopleSoft employees, and the catastrophic fallout anticipated as part of that merger never happened; with the current economic situation, it probably won’t happen this time either. I also have to say this is a much better fit, with Oracle being the acquirer, than it would have been with IBM or HP. The product lines are more complimentary than IBM’s or HP’s, and I suspect there will be fewer layoffs than if either of those companies had made the acquisition. Sun’s people may not like the culture, but I have been hearing complaints from current and ex-Sun employees for years that they were unable to win market share despite having really innovative technologies, and there will be a sense of pride in having the products you worked on effectively marketed and sold.

When I worked at Oracle way back when, it was amazing to watch the sales dynamic that was going on. If the customer was making a $20M purchase of hardware and software, let’s say $17M of that was for the hardware. However, the customer’s motivation for the purchase was they needed a solid database platform. That meant the $3M Oracle purchase is what mattered to the customer, and how well Oracle performed on the hardware was the deciding factor in the purchase. This meant the smaller database software company held sway over the larger hardware vendors. For years Oracle has used this incredible leverage over their hardware partners and ‘squeezed’ them on pricing. Now Oracle is the huge company with great margins, but the market dynamic is really changing, and commoditization is moving right up the stack and squeezing their core business as well. It’s not just about the database any longer.

Look no further than Cisco getting into the Server/Switch business and offering a unique take on virtualization and provisioning. Several people I spoke with at the RSA conference all said the same thing: Oracle needs to own more of the data center in the coming years if they want to continue their growth curve. I believe Mr. Ellison meant “We’ll engineer the Oracle database and Solaris operating system together. With Sun we can make all components of the IT stack integrated and work well.” quite literally, and it reflects Oracle’s long-term growth strategy. Bundling Solaris with whatever virtualization technologies are at their disposal, InfiniBand Switches, and a full array of servers, gives Oracle a chip-to-web-app presence in the data center that makes the LAMP stack look like a child’s toy.

From a security perspective, Oracle now has some really compelling technologies at their disposal. Trusted Solaris is the most secure general purpose OS in the world. Sun’s data encryption and authentication/key management may not be best of breed, but they are solid products that could generate considerable revenue in the hands of Oracle’s professional service arm. And while it is really difficult to secure a JVM properly, it can be done, and the beauty of the Java programming language is that it flat out has the best object model I have ever used. I can properly encapsulate and protect objects, and the language syntax is far easier to read and analyze for coding and security flaws than C++ or other commonly used environments. If Oracle decides to knit these components together within their Data Vault variant of the Oracle database, you will have all of the elements for a very secure development environment.

One of the rumors that I was hearing was that Oracle would kill off MySQL. This has been covered in some of the blogs as well. I personally think this is nonsense. MySQL is a very well-designed database. It is modular and cannot only be tuned like an Oracle database, but is instead configured more like a Linux kernel to meet the user’s specific needs. MySQL has a rabid following and what I am estimate at around 15 million installations around the world. When you couple this with the BEA pieces in place and the Java programming language and associated tools/platforms Sun has, you have a really phenomenal web application development suite. Oracle no longer has to ‘compete’ with MySQL – now they have a real answer to PostgreSQL (No, Oracle Lite fans, that was not the answer) without undermining their core database business. What Oracle really needs to do is provide a PL/SQL parser/pre-processor for MySQL, thus providing developers not only the option to use existing SQL/PSM, but the Oracle-specific procedural language most DBAs are familiar with. This would keep the existing MySQL users happy, and offer a migration path into the core Oracle database platform should they outgrow MySQL’s capabilities.

Also keep in mind that Oracle purchased Innobase InnoDB, which is not really a database, but rather an underlying storage engine that is commonly used by MySQL. One of the cool things about MySQL is that you can configure it with different storage backends, such as clustering or ISAM. So Oracle owns MySQL and one of the commonly used storage technologies for it, and that platform has strong user affinity – now they just need to find a way to leverage that and make money from it. Letting that community wither and die just does not make sense.

To me this looks like a very complimentary acquisition.

–Adrian Lane