Lead

Can Oracle successfully assimilate Sun?

Oracle finds itself on a sticky wicket after gobbling up Sun Microsystems. While the rationale of the deal is quite clear, Oracle bought Sun for Java and Solaris, questions remain vis-à-vis MySQL and Sun’s hardware business. Akhtar Pasha writes that this mammoth acquisition could prove to be Oracle’s toughest deal yet

Oracle’s announcement that it would acquire Sun took many industry observers by surprise. The database giant has offered little detail on how it intends to achieve the $1.5 billion in first-year operating profit that it has promised investors. While Oracle’s interest in Java and Solaris has been widely discussed and publicized, the company declined to participate in this story citing the fact that the deal is still pending approval. Even otherwise, Oracle has not been forthcoming on critical aspects of the deal, namely, the future of MySQL and what it intends to do about Sun’s hardware business. Customers, analysts and the financial community all have their own doubts at this point of time.

MySQL’s role is unclear

Will Oracle kill MySQL or slowly fold it into its own database? Analysts and customers are asking this billion-dollar question. The fact is that Oracle will gain in the database space from owning MySQL. Sun had a huge SMB base for this product and a steady revenue stream of about 3% of its revenues accrued from this product. This is an opportunity for Oracle to garner additional revenue from the subscription license base but the company needs to position MySQL clearly in the context of its existing database offerings. MySQL will give Oracle a standing to take on Microsoft in the SMB space where the latter is strong.

Ray Wang, Vice President, Forrester Research, said, “Access to MySQL provides a key entry-point into the SMB market. If Oracle executes well, MySQL could be the icing on the cake because Oracle had struggled in penetrating the SMB database market, which is where Microsoft SQL Server has dominated for years.”

He added “MySQL has achieved high adoption [as the backend database for] small to moderately sized Web-based applications and is extensively used by universities and SMBs; therefore, killing MySQL would be a huge mistake.”

We expect Oracle to invest in MySQL in an attempt to trump Microsoft SQL Server in the SMB segment where Oracle’s flagship database has typically has been either too complex to manage and deploy or too expensive to implement. Forrester expects that Oracle will improve the product’s usability, auditing, encryption, and monitoring and create a seamless migration path from MySQL to Oracle DBMS in order to help MySQL play this vital role. If MySQL fails in the SMB market, Oracle will relegate it to the developer tool bin along with its other open source database products such as Berkeley DB and InnoDB.

Another unclear aspect is Oracle’s commitment to MySQL’s dual-licensing model (free and paid licenses).

That being said, a portfolio combining MySQL and Oracle database management systems (DBMS’) could make Oracle the most powerful database company that covers all types of applications and databases for small to large enterprises.

As every vendor purports openness, some things such as tuning, migration, and integration may work better with Oracle. Wang said, “Expect Oracle to roll out a database machine—creating a stack play. Larry Ellison has always promoted the idea of a database machine, going back to the early 1990s when he invested in a company called nCUBE. Although adoption of the nCUBE MPP database machine never took off, (largely because databases were relatively smaller in size and the cost was prohibitive), Oracle’s acquisition of Sun could change the game.”

Additionally, Oracle made a similar move last year by rolling out the HP Oracle Database Machine, a system using HP hardware and storage, which has had some success. “Oracle has already demonstrated strong engineering on operating systems, databases, and file systems, so acquiring Sun servers and storage puts a database machine well within its reach. With enterprises struggling with growing data volumes and increasing numbers of databases, the business case for a database machine becomes more compelling to lower cost, improve IT productivity, and increase performance scalability,” added Wang.

Competitors waiting to pounce

Certainly, the competition is not sitting still and it sees a window of opportunity to poach Sun’s existing customers. Pallavi Kathuria, Director, Server Business Group, Microsoft India, said, “At this point we cannot speculate how Oracle’s acquisition of Sun will rationalize their product roadmap regarding MySQL or any product in their portfolio or strategy moving forward. We can consider however, the integration of competing products, or new products within the portfolio, will potentially lead to great complexity for customers.” She added that integrated solutions that work well together, offer business value and are delivered at a competitive price point are important to customers.

Kathuria added, “With each acquisition, Oracle increases the complexity of its stack with competing technologies and, therefore, introduces uncertainty and risk for their customers. Additionally, this acquisition raises questions about the future of Java and style of engagement that Oracle will pursue with the open source community. We hope that Oracle will continue promoting the openness of Java on Windows as Sun has done in the past.”

Asheesh Raina, Principal Research Analyst, Gartner sees this Oracle-Sun synergy in a different light. According to him, “We need to understand that Oracle and Sun are bringing two divergent technologies and architectures under one roof and anything new from Oracle would take 12-24 months [to deliver] as Sun products have to be seamlessly integrated into Oracle’s portfolio.”

Raina continued, “We do not foresee a forced migration of MySQL into Oracle database. We do anticipate aggressive support by Oracle failing which customers could move to IBM DB2; IBM is getting aggressive [on this front].”

Until now, Oracle and Microsoft SQL Server have both been gaining market share (often at the cost of IBM DB2, open source databases etc.). To address this market opportunity IBM has formed partnership with EnterpriseDB to embed Enterprise DB Postgres server into IBM DB2 9.7 version code named as Cobra. IBM feels that this sandwich strategy would help it win back customers.

V Subramanyam, Vice President, Information Management, Software Group, IBM India/South Asia, said, “To allow an application written for one RDBMS to run on another virtually unchanged, and many pieces have to fall into place. Different locking mechanisms, data-types, SQL, procedural language residing on the server, and even the client interfaces used by the application itself need to be aligned not only in syntax, but also in semantics.”

He continued that IBM DB2 9.7 uses the same native PL/SQL that Oracle’s eponymous database uses. Further IBM has started supporting stored procedures, triggers and Oracle datasets into DB2 9.7. Subramanyam added, “This allows many applications written against Oracle to execute against DB2. It [IBM DB2 9.7] removes all the obstacles that prevent Oracle customers from adopting DB2 and it dramatically lower the barriers for applications written for Oracle by enabling them on DB2. The intention is to make it as easy as possible for them to adopt DB2, whether their applications were originally written for DB2 or not.”

Hardware’s a knotty problem

Previously, Oracle only had middleware and enterprise applications to sell. Buying Sun gives it an extensive hardware portfolio, both on the server side with SPARC, CMT and x86 gear and on the client side with Sun Ray thin clients. Along with the ERP applications that Oracle already has, this means a business could buy everything from one vendor. Nevertheless, this is easier said than done.

Oracle has a reputation for hard-nosed cost cutting in the companies that it acquires; it has already announced that it plans to reap $1.5 billion in operating profit from Sun in the first year, and more than $2 billion in the second. How it intends to achieve this considering that Sun had posted losses in each of the last three quarters in a row is something that it has yet to state. Oracle’s goal for operating profit and Sun’s operating loss are both facts and the two don’t equate.

Further, Oracle’s commitment to its shareholders that it will get $1.5 billion in operating profit from Sun in the first years has put it in a precarious position. Firstly, there is no question of not meeting this commitment to shareholders because if it doesn’t, then investors will lose confidence in the company. Secondly, we feel that Oracle can achieve this target in one of two ways. The first would be to boost Sun’s revenues by jacking up the hardware and maintenance costs for Sun’s existing enterprise customers, which might come at the cost of losing some in the short term. Shailesh Agarwal, Vice President-Sales Programs, IBM STG - India/SA, said, “It will be interesting to see as how Oracle meets its stated profit commitments. To achieve such a dramatic turnaround in the short term may require a drastic increase in prices for Sun customers especially around support services. Alternatively, it may mean a drastic reduction in costs—not just administrative costs but also a significant part of Sun’s $1.8 billion R&D spends. Either of the actions may be bad news for Sun customers.”

Further, these stark numbers underscore the fact that Oracle has major cost cutting slated for its new acquisition. While we expect many of the cuts, which would be significant to aim at Sun’s sales and administrative expenses, the company’s hardware business may also be trimmed including R&D as pointed out by Agarwal. Additionally Sun’s SPARC processor business, for instance, has been a major focus of the company’s $1.8 billion annual R&D budget. However, this may be axed or that portion of the business sold off by Oracle, which is likely more interested in Sun’s cheaper Intel- and AMD-based servers. The company’s tape storage business, a result of its $4.1 billion acquisition of StorageTek in 2005, has even less synergy with Oracle’s business.

Considering that Oracle is unfamiliar with the business dynamics of the hardware business, logistics and channels involved and the operating margins in hardware are wafer thin, it does not mean that Oracle will sell off the hardware business immediately, at least not for two to three years. It will likely keep Sun’s server business, which pulls in hefty maintenance revenues. About 40% of Sun’s $13.9 billion revenue, as of AMJ 2009, comes from service relationships. Raina said, “Oracle was never in the hardware business and hence it would face a big challenge in managing the wafer thin margins and channel partners. We think that Oracle will come out with pre-integrated offers for the SMB market leveraging Sun's hardware business. Additionally the hardware business is important for Oracle's strategy on virtualization and cloud computing that needs a solid IT infrastructure riding on servers, storage and networking.” Industry experts feel that Oracle will take the appliance route for delivering its applications in a pre-packaged bundle.

IBM sees a substantial opportunity in providing risk mitigation and migration paths to Sun’s customers. Although IBM and Oracle are going to compete in different domains, they are also partners in certain areas so there is no question of conflict between them.

Analysts and customers feel that while hardware cannot be Oracle’s core business, nevertheless, servers and storage are the nuclei of most infrastructure and Sun customers would be unhappy about the fact that they would be forced to cough up additional amounts for maintenance directly adding to Oracle’s bottom line.

Agarwal said, “It is hard to see servers or storage being a core business for Oracle. We have already seen customers exploring alternatives. IBM is positioned to provide a stable alternative.” He continued that IBM had invested significantly around migration services to provide a smooth transition. “We are conducting workshops showcasing the migration route for Sun’s customers to IBM’s platform. We are helping them in doing cost and effectiveness assessments of migration to make a business case without increasing their’ operational costs. With help of IBM Global finance (IGF), we are in position to provide IBM infrastructure in an OPEX model. We find that, in most cases, the OPEX cost of new IBM servers is not too different from the current operational costs of old Sun servers. Customers can thus achieve migration without any change in cash flows.”

Solaris into Linux

According to analysts, Solaris has assets [features] that would be quite useful on Linux and if Oracle were inclined to use these in its Unbreakable Linux offering, it could change the licensing model and make it possible to integrate Solaris assets back into Linux. For instance, developers have been eyeing Sun’s DTrace technology for possible use in Linux. The DTrace tool helps systems administrators diagnose and troubleshoot problems in applications and the operating system kernel in real-time. The Containers technology in Solaris would be useful in Linux as well as it allows the OS to be partitioned and run instances independently. However, it is unlikely that Oracle would get rid of the commercial version of Solaris in favor of Linux, primarily because of the steady stream of maintenance revenue that Sun derives from this popular server OS. Both Sun and Oracle get professional services revenues from Solaris—typically, it’s a 90%-plus margin business [as per an Ovum analyst] and Oracle is unlikely to do anything to damage that extremely profitable maintenance/ support business.

What is required from Oracle is to demonstrate and communicate to its customers how it plans to support Sun customers from here on and what strategies it has put together to achieve the $1.5 billion in profit in the first year of operations from Sun’s business.

akhtar.pasha@expressindia.com