Original URL: http://www.theregister.co.uk/2009/04/21/oracle_sun_open_source/

Sun and Oracle: End of a beautiful dream

Open source goes to work

By Gavin Clarke

Posted in Business, 21st April 2009 00:20 GMT

Sun Microsystems' soon-to-be ex-chief executive has been painting his company's acquisition by database giant Oracle in positive tones.

Jonathan Schwartz called the proposed $5.6bn deal a "fantastic day for Sun's customers, developers, partners and employees across the globe". It will also "redefine" important boundaries in IT by eliminating cost and complexity in an industry focused on components.

At least, that's what he appeared to be saying.

The only thing that's going to get redefined is everything that Schwartz and Sun have worked towards for five-plus years: platform independence, open systems, and affordability. but Sun was in no position to implement these grand ideas thanks to its inertia and bureaucracy

Unless there's a radical shift in strategic direction or management thinking at Oracle, and unless Oracle adopts Sun's woolly and academic ways of thinking and operating, what you're looking at is the death of the philosophy of benevolent big-company sponsorship of Java and open-source, and the death of big-company price competition.

Some have said that Oracle is a better fit than IBM would have been, because of the clash between IBM's stuffed-shirted East Coast mindset and the free-thinking experimentalism of Sun. But Oracle is not a dream marriage. Oracle is a tough place. It routinely evaluates the performance of staff and chops around 10 per cent. Acquired staff and products must also justify their place in this Dawnian environment or they'll get chopped or boxed.

For example, Oracle spiked its own Java application server having bought the rival WebLogic Server from BEA Systems, because WebLogic was the better application server. It canned development of BEA's WebLogic Workshop development environment.

Oracle will take the decisions Sun could not - and that's what'll have people at Sun worried. And while change should be welcomed, there's no way this should be seen as a bright new dawn for Sun customers or those who've come to believe in its actions on open source or Java.

The first thing you can expect from a Oracle acquisition is due diligence of the assets and a comparative analysis where Oracle has competing assets. Oracle will weigh up what's worth keeping and jettison the rest. The latter will be marked by end-of-lifing via support and maintenance, or releasing code to the community - where it will fade and die.

Oracle initially had teamed up with Hewlett-Packard to buy just Sun's software business - or, rather, parts of it: Java, Solaris and MySQL. That deal was blocked by IBM.

These three software assets remain at the heart of this deal, only this time Oracle is also getting Sun's server and storage business for an additional $3.6bn.

So, why Java? Java is strategically critical to Oracle's middleware, development tools, application server, and database support. Oracle has been a life-long believer in Java, which - running on Linux - provides the perfect answer to Windows and .NET from Microsoft. Java also means licensing from all those enterprise ISVs and device manufacturers.

Solaris is important to Oracle, too. More copies of Oracle's database run on Solaris than on any another operating system - a fact chief executive Larry Ellison helpfully pointed out on Monday. Ellison talked of tuning its database to Solaris, which will help keep Oracle in major accounts such as telcos, service providers, and banks and financial services that might otherwise have been flirting with Linux and its database or - worse - Linux and MySQL and Postgres.

Whither MySQL?

MySQL is more complicated, but it provides a ready-made opportunity for Oracle ISV and systems-integrator partners.

Many will be worried about the future of MySQL under Oracle - understandably so. but the deal likely means the continuation of development of the database. Oracle kept InnoDB going, despite buying it and removing a major pillar from beneath MySQL. It has also kept Sleepycat going. MySQL will give Oracle a bigger stake in the OEM, ISV, and web-development business - especially given that more than half of web developers use the database. The chafing will occur when you see customers who want to deploy MySQL in the enterprise instead of, say, Oracle.

Killing MySQL makes no sense. In fact, you might just see MySQL thrive under a leadership actually geared to making some money from software rather than just loving the idea of software and smothering it to death.

Sun has got a lot of other open-source software work. I've written a lot about the shortcomings of Sun's open-source software, but - surprisingly - this deal does not represent an automatic death sentence of all of these projects. This is where Oracle's due-diligence will kick in and where you can kiss good-bye to large-scale independent open source.

Optimists inside Sun will likely argue GlassFish, OpenSolaris, and Java CAPs can serve their new master in the same OEM market using MySQL. However, this would assume Oracle is interested in building an open, platform-neutral, and affordable third-party partner community. It isn't.

Oracle wants partners building on Oracle software that it can charge for, and that partners and customers must keep coming back to Oracle for, in order to expand its core business.

The biggest determining factor, though, will be cash generation. If any of these products were making money, they would pass Oracle's acid test for survival. They aren't.

You can also finally say goodbye to NetBeans, Sun's open-source Java-tools platform alternative to Eclipse. Oracle and Sun had a very public falling out in January 2006, when Oracle's head of server technologies said that his company had no plans to adopt NetBeans or any of its technologies - which was contrary to statements made by Schwartz. Oracle is Eclipse and its JDeveloper integrated development down straight down the line. There is no room for NetBeans.

Sun's open source will live on when it comes to building that integrated Sun-Oracle appliance that Oracle's president Charles Phillips mentioned while announcing the acquisition on Monday morning. Oracle is likely to build this using a combination of software that its engineers have judged valuable following careful evaluation. It will use this to build more server appliances following on from the HP Oracle Exadata Storage Server and HP Oracle Database Machine unveiled last October.

There can be little doubt that ZFS and D-Trace in Solaris and OpenSolaris are superior technologies. These will have tickets to ride in any Oracle-based systems that serve really large, information-serving and storing hardware and software systems from Oracle. And committed to such systems, Oracle is. The idea of a server appliance is one of Ellison's most cherished dreams, starting in 1998 with the 8i Database Appliance.

The idea of the appliance is an evolution of Oracle's earlier belief in the software stack - application server, database, middleware, and applications - all coming from the same provider. That Schwartz should now endorse this is not a surprise, but it is a shocking U-turn on his work at Sun.

Component elements of Sun's open-source software such as ZFS and OpenSolaris, combined with Sun's storage and Solaris, mean big bucks for Oracle. The company's underlying strategy is as an "information company". Hardware and software, be they appliances that speed up access to customers' information or software that makes systems mission-critical and cheaper to build, will survive the due diligence that's coming.

Whither hardware?

The only uncertain part here is how far Oracle really wants to become a hardware company. Despite all that bravado of Oracle becoming a hardware company with last year's Exadata and database machine, the devices are actually built by HP.

For Oracle to become a genuine hardware business would be like getting a super tanker to change course. Oracle is a software business. Oracle could run Sun's systems business to manufacture those Exadata and database machines, and build bigger and better database and storage systems for customers such as telcos and banks.

Oracle president Safra Catz promised on Monday that Oracle would turn Sun's systems business into a profitable operating unit within Oracle. But long-term? Has Oracle really become a hardware company? Why not just continue to partner with HP, especially since hardware is a notoriously less profitable business than software? The strategy must be to return Sun's systems business to profitability - but only to then sell it at a greater profit.

The final casualty of this deal will be some of the daring thinking that Schwartz and Sun introduced to enterprise software in the early days of his tenure as senior vice president for Sun's software business. Sun adopted ideas of paying by company or individual on a subscription basis, to make big systems more affordable and - gasp - fairer.

OK, Sun was pushing software nobody really wanted and coming from behind in a middleware market it had already lost - it had to offer something new. But such ideas got people thinking, and helped contribute to fresh approaches to charging for enterprise infrastructure that you see today.

Such ideas have no place at Oracle. One of the first things that Oracle does when it acquires a company is increase the price of that company's software. Customers of MySQL should look out, as should - incidentally - licensees of Java.

Oracle is a company many in Silcon Valley simultaneously admire and hate. It's seen as a success because of the way it's run. But the way it's run is seen as arrogant and a blot on both independence and innovation. The focus on process and pursuit of growth has turned Oracle into a company that buys and integrates other people's technologies into the non-stop express that is Oracle.

Unless there's been a radical shift of thinking at Oracle - a shift away from pure profit and away from performance and process and towards experimentalism - then you will see a radical shake-up for Sun and its products. For some products and units this will mean improved management and the ability to make money for their owner. For others, it'll be the end of the line.

It'll be hard, too, to see where those fellows and big thinkers such as Java father James Gosling, XML co-inventor and languages expert Tim Bray, or crypto authority Whitfield Diffie will fit. Sun's academic culture promoted tinkering, thinking, and experimenting. Oracle is business first and focused on integrating a lot of moving parts, and not on innovating or breaking new ground in computing, languages, or security. Again, unless there's a change, the big thinkers that Sun prized as its brain trust are likely to take their leave.

but the biggest casualty will be the loss of a corporate-sized backer of open source that - thanks to its failures in middleware and history in creating Java - advocated open systems over traditional vendor lock-in.

Sun liked to brag about how much code it had contributed to open source over the years. Oracle, though, has never made such lofty claims. Instead, it's used open source to advance its business or to try to close down the competition.

Open source will continue at Oracle - along with Java. It could even profit. Just don't expect it to help anybody else. ®