Burton Group on what Oracle will do to Sun and what it means to users

With Oracle's acquisition of Sun imminent, Sun users are surely asking what the software company plans to do with its new toys. Burton Group Analyst Richard Jones explores scenarios and offers advice on how users can respond to changes.

All is quiet while Oracle and Sun wait for their merger.

But Sun customers sit uneasy, wondering what long-term implications Oracle's roadmap will have on their Solaris, Sun SPARC rackmount servers and blades, Sun AMD- and Intel-based rackmount servers and blades, Sun storage, Sun StorageTek tape solutions, and Sunray investments. Until Oracle's plans unfold -- most likely over the next year -- we can only speculate on what will transpire. However, Oracle's past actions and business focus can give us some good hints as to what will most likely happen. First off, we need to understand Oracle's main focus, and then we can determine the best course of action to take with Sun investments.

Oracle's business revolves around the database

For Oracle, the world revolves around the database. The database is at the center, with line-of-business applications moving up the stack and operating systems plus hardware moving down the stack. In reviewing the major "up the stack" acquisitions that Oracle has made over the past few years, we can see that each has focused on increasing the value of the database and building additional revenue for the company. JDEdwards, Siebel and Peoplesoft are all examples of critical line-of-business applications that depend on the database, underscoring its importance to the business world. Oracle has not marginalized these line-of-business applications but has taken "the whole is greater than the sum of the parts" approach for the applications and the database put together. Furthermore, Oracle has built out its Fusion Middleware framework to allow business-specific customization of its line-of-business applications. This framework utilizes the Sun Java J2EE platform.

However, looking "down the stack," Oracle has taken the approach of commoditizing everything below the database -- reducing the cost of the OS and hardware to commodity levels to increase the value of components above the operating system. So what has Oracle done over the past few years in this regard? In 2006, Oracle launched Oracle Enterprise Linux (OEL) as the preferred platform for the latest Oracle database. The 10 G version of the database was developed and tested on Linux prior to porting and testing on other platforms such as Solaris, Windows and HP-UX. This was a change from the past, where the prior database releases were first developed and tested on Solaris before being ported and tested on additional platforms, including Linux. Oracle began its Linux-on-x86-commodity-hardware focus in earnest as a way to reduce the total solution cost to their customers while maintaining the profit margins on the database and higher products.

The Oracle Unbreakable Linux product offers enterprise Linux subscription support at a fraction of the cost of Red Hat Enterprise Linux (RHEL) Advanced edition. OEL is essentially the same code as RHEL, with the trademarks and logos changed from Red Hat to Oracle. This is what open source is all about -- distribution vendors compete on the service offerings they provide, not the code, because the code is freely available to anyone under the GNU General Public License.

So applying these goals to the Sun acquisition, we can get an idea of what is likely to occur. First, we need to look at all that Oracle gets when it acquires Sun.

Oracle's Sun inventory

Sun Microsystems brings technologies from hardware and storage up through end-user applications to the Oracle product family. These include the following general components (those in bold overlap with Oracle technologies):

  • Star Office -- An end-user office productivity suite that competes with Microsoft Office. The open source version is called OpenOffice.
  • MySQL -- An open source database targeted at small and medium-sized business (SMB) applications
  • Java platforms -- A application development platform environment used widely for Web services applications
  • Solaris -- An enterprise Unix-based operating system family available for both Sun's SPARC-based server machines and commodity x86 platforms
  • xVM -- An x86-based virtual machine hypervisor built on Xen open source technology
  • Sun SPARC servers and blades -- A RISC-based enterprise server hardware platform based on Sun's own processor architecture called SPARC
  • Sun Fire x86 series servers and blades -- An x86-based enterprise server hardware platform based on AMD's and Intel's popular x86 processors
  • Sun Ray -- A custom-built thin-client terminal and associated software for thin-client deployments
  • Sun Storage -- A family of disk- and, more recently, SSD-based storage array platforms
  • StorageTek -- A family of tape-based data archiving hardware focused on large tape libraries.

At first blush, one would think that MySQL, Solaris and xVM would be prime candidates for the chopping block after the acquisition. However, the situation is more complex. Oracle originally only wanted Sun's software business unit, and was not interested in the server, thin-client and storage hardware business. Furthermore, the steep decline in Sun's hardware business around its proprietary SPARC processor is what put Sun into its current predicament. So what is Oracle likely to do?

Oracle's probable moves, post-Sun acquisition

Oracle will not make any sudden moves that would jeopardize its current database and line-of-business applications customers. Many enterprise Oracle database deployments are still on Solaris SPARC hardware. Oracle must keep those customers happy. And enterprise deployments are slow to move, so Oracle will need to keep the Solaris/SPARC combination supported for nearly a decade, even if new system sales are discontinued soon after the acquisition completes.

Competition is also driving Oracle to take the moves slowly. Even prior to the completion of the deal, rival IBM is swooping in to steal away Sun customers -- a move that threatens Oracle's core business, as IBM's next step would be to entice those customers to move to its DB2 database from Oracle's flagship product.

Longer term, however, Oracle will discontinue the SPARC server line, selling those assets to a suitor such as Hitachi, a current SPARC OEM. The storage business will also be sold, probably to a foreign suitor. The Sun x86 server line will most likely be sold as well so Oracle doesn't remain in competition with its hardware partners Hewlett-Packard, Dell and IBM.

Sun Ray doesn't fit in Oracle's data center focus, and will most likely be sold.

Sun's recent moves to position Solaris as an open source alternative to Linux can give Solaris a longer life, but expect Oracle to take key Solaris technologies and work them into Linux with a longer-term goal of moving Sun customers to Linux and giving Solaris to the open source community to continue forward. Universities and large corporations stuck on ancient applications that require Solaris would keep it alive until no one needed it any longer.

Oracle will move Java forward. The Java platform is the Sun asset that is the most important to Oracle's core business.

MySQL will be an interesting decision for Oracle, and I can imagine a number of arguments within the halls of Oracle over what to do with the product. Oracle would be best served by positioning MySQL down market to SMBs to compete with lower-end deployments of Microsoft SQL server. This move would be difficult for Oracle, however, as Oracle lacks the reseller and distribution channels into that market. I view this as an opportunity for Oracle that it could build out to expand market reach.

StarOffice will most likely be rolled back into the OpenOffice project and continued within the open source community, where it has seen broad support from all the major Linux distribution vendors.

What Sun customers should do?

The wheels were already in motion with respect to Sun's SPARC-based hardware business prior to the Oracle acquisition. If you are on SPARC, your plans should already be in place to move to x86 or IBM's RISC platforms. Concerns about x86 scalability and performance are now old history with Intel's recent launch of the Xeon 5500 (Nehalem) chip. The Nehalem CPU marks a significant performance and scale-up improvement in the x86 processor line.

Solaris users will have more time before they have to worry about moving operating platforms. However, if you had plans in place or were already considering a migration to Linux, you should continue with those plans. Moving from Solaris to Linux should be in your long-range plans, such as five years out or in conjunction with your next full hardware refresh cycle. Expect to see the features you like in Solaris to slowly appear in Linux over time. They may not be exactly the same, but will be good enough.

Sun storage users should wait and see what Oracle does with that business unit over the next year. You should be safe through your current investment's lifecycle, but a contingency plan to move to a competitive storage vendor would be wise.

In a future article, I will discuss the issues surrounding the x86 hypervisors that Oracle has acquired. After the Sun acquisition is complete, Oracle will have three Xen-based hypervisors: OVM, xVM and Virtual Iron.

The bottom line of the Oracle-Sun deal

Sun customers can expect acceleration to end of life for those Sun products that were already losing market share, such as SPARC-based hardware. Oracle will cater to its database and line-of-business applications customers that are also using Sun products, but moves to Oracle's Linux operating system and hardware partners will occur over time.

Read more on IT efficiency and sustainability