Oracle's recent announcement that it will acquire Sun is such a major shift in the industry landscape that it's hard to find an area of the industry that it does not affect. It seems like only yesterday that Sun was part of a coterie proclaiming the death of the mainframe and offering migration programs.
A quick search of the Sun website shows several "mainframe rehosting" press releases from the past year. On the other hand, neither Oracle nor Sun mentioned the mainframe in their joint announcement teleconference. Just what effects will Oracle's acquisition of Sun have on IBM mainframe sales and mainframe customers?
To understand those likely effects, it's useful to look at the universe from Oracle's point of view. In my opinion, Oracle sees Sun as primarily of use not for its hardware, but for its infrastructure software. Oracle's database and middleware segment is on track for a $12B fiscal year, almost double what Oracle is paying for Sun (estimated at $7.4 billion).
In constant-currency terms, Oracle's infrastructure software business has continued to grow in the last three quarters, unlike Oracle's application business and Sun's hardware business. Thus, Sun's value-add in software and, via differentiation software, in services (i.e., customized solutions) matters far more to Oracle than hardware revenues.
A brief summary of Sun's value-add to Oracle in infrastructure software goes like this:
- one-stop rather than a two-stop shop for (Java) development needs;
- Development features in Java tools that support scalability to the very highest levels, like in analytics;
- A possible market niche in embedded-software development;
- An integrated storage/software stack to allow better performance and scalability in data-intensive, mission-critical applications;
- Ability to rapidly customize and offer a full one-stop solution for large enterprise data centers;
- A low-end business process management and composite application development tool (Java CAPS) to complement its high-end ones.
Note that the flashier aspects of Sun, such as Solaris, SPARC, Sun Storage, MySQL and Glassfish, are not as important to Oracle as is the above list.
When one views the Sun value-add list from the point of view of mainframes, there is no direct impact on the mainframe market. Java is already available on the mainframe, and it is unlikely that Oracle will choke off its availability. The mainframe already has (to a certain extent) an integrated storage/database offering, and will still have a one-stop enterprise data center solution that is more comprehensive than Oracle's -- although the ability to rapidly customize is still a work in progress. And we may note that "mainframe rehosting" has already been effectively rebutted as a source of total cost of ownership reductions.
My conclusion is that the effects of Oracle's Sun acquisition on the mainframe world are very likely to be not only indirect, but also long-term. What are those effects likely to be?
Long-term effects on the mainframe
The first effect that mainframe users should consider is the further integration of storage area networks (SANs) and information lifecycle management (ILM) with databases such as Oracle Database and IBM DB2.This means that the two sets of software would begin to share metadata about data storage patterns (ensuring that storage is optimized for both the individual application and the overall operational-app portfolio) and data "age" (ensuring that new data that's likely to be changed is put on a higher-performance disk, and speeding backup/recovery). My studies suggest that effective coordination of storage and database can improve database performance by more than 10%, and effective ILM can cut backup/recovery time in half. It is likely that in the long run, IBM will match Oracle's efforts in this area -- a major benefit to enterprise mainframe customers whose enterprise-app scalability is crucial to business success.
The second effect is that in competitive IT and business services bidding, Oracle plus Sun will be a much more credible alternative than either Oracle or Sun alone. However, the field of competition is now shifting, as users now look not only to reduce TCO in tough economic times but also to flexibly reduce costs by abstracting virtual machines to run remotely in a cloud -- a new variant of capacity on demand. It appears that in mainframe cloud support, IBM is well positioned to offer leadership solutions in the foreseeable future. So Oracle's new service prowess may well affect the share of new market that the mainframe takes; but not the share taken by growing existing installations for new needs.
The third effect is the shift of Sun within Oracle from a hardware to a services sell. No matter what Sun's announced strategy was, the fact of the matter is that the bulk of revenues continued to come from selling hardware, not services. By contrast, Sun within Oracle is effectively a services-led sell. In effect, Sun will finally join Hewlett-Packard Co. and IBM in treating hardware as more or less a commodity. So the mainframe's competitors for market share will indeed get stronger.
The negative effect of the stronger competition, however, is likely to be minimal -- as long as IBM continues its mainframe marketing strategy of promoting the mainframe as a hub. The hub strategy is an "undermine from within" one: It says that in the long run, the mainframe will be the best machine for some tasks, and other platforms for others. This de-links the service sell from the hardware decision and allows the mainframe to compete using its strengths -- scalability, robustness, security -- so that as it takes over some tasks in the data center, it qualifies for new share of market as those tasks evolve. This way of incrementally growing task responsibilities and using those to drive new sales is impervious to "I can offer you a full one-stop shop" arguments in an era of heterogeneous solutions.
It is tempting to see Oracle's acquisition of Sun as a morality play, in which either the noble Oracle folks allow Sun to fulfill its full potential as a new technology mainframe killer, or the evil Oracle folks deep-six the SPARC/Solaris technology in search of short-term strip-and-run savings. By all indications, neither will be true. Instead, existing Oracle and Sun customers will see immediate benefits -- one-stop Sun/Java plus Oracle Database shopping, to be followed before the end of the year by Oracle plus Sun/Java/SQL programming and the first performance fruits of Oracle Database-Sun storage integration -- but these will not impact the mainframe market in any significant way.
In the longer run, customers in particular and the industry in general should look to Oracle plus Sun for new high-water marks in database performance, better integration between open source and Oracle databases, and more scalable Java programs. Again, however, these should not negatively affect the mainframe in any significant way, as long as IBM continues its hub marketing strategy.
No merger this major is without its effects in every aspect of computing. Still, you should feel comfortable in concluding that the only merger-caused surprises upcoming for the mainframe and its customers are very likely to be positive ones. The world without Sun to brashly take on all comers and criticize an IBM-Apple alliance as "purple applesauce" will indeed be duller; but the pace of improvement in the mainframe should continue unabated -- as it has for almost 50 years.
ABOUT THE AUTHOR: Wayne Kernochan is president of Infostructure Associates, an affiliate of Valley View Ventures. Infostructure Associates aims to provide thought leadership and sound advice to vendors and users of information technology. This document is the result of Infostructure Associates-sponsored research. Infostructure Associates believes that its findings are objective and represent the best analysis available at the time of publication.
What did you think of this feature? Write to SearchDataCenter.com's Matt Stansberry about your data center concerns at email@example.com.