There are a number of companies in different phases of making the switch from Sun/Oracle to IBM and DB2 (IBM has a white paper out on this), but iHealth is the first one that I've been able to interview that is far enough along in the process to have something meaningful to share. Its migration isn't complete but it has passed initial trials and the hardware is already in place. Clearly the hardest part-the database migration-is yet to come, so we'll need to revisit this in a few months to get the full story.
This is what I learned.
iHealth Under Sun/Oracle
iHealth is a company that processes insurance claims. While it is privately held, it is also an enterprise-class company. Before the migration, iHealth was heavily invested in Oracle Sun with two large, fully-loaded Sun E25Ks and 250 additional Sun servers of various sizes representing a decade of significant investments in Sun and Oracle technology. It also employed six DB 6 professional admins. However, like most companies that decide a major change is needed, iHealth had become unhappy with Sun's roadmap as Sun went through the Oracle acquisition process and it was very concerned that Oracle software maintenance charges were increasing at high rates and couldn't be guaranteed for long enough periods of time.
iHealth was particularly annoyed by Oracle's practice of charging penalties for cores that weren't in the same box given the administrative benefits for technologies like virtualization. And even though iHealth was increasingly dissatisfied, Oracle continued to send out the "Flunky of the Week" to address issues. iHealth felt that Oracle failed to take its concerns seriously.
It was time for a change.
HP's Failed Bid
HP was the other bidder for iHealth's business but iHealth wasn't comfortable with HP's roadmap and commitment to large-scale servers. HP has a lot of products, but was unable to resolve iHealth's concerns with regard to product direction for what was going to be a massive purchase. iHealth also felt that AIX was closer to Solaris than HP was, making the migration less risky.
iHealth Under IBM
iHealth was able to replace the two Sun E25K servers and all of the 250 miscellaneous Sun servers with two large IBM I95 P6s with enough headroom left for substantial growth. Because of the long approval time and trial, it chose not to go with the newer P7s or wait for the soon-to-be-released P8s. These servers have already been deployed and the initial hardware migration went virtually seamlessly with just a few minor problems. IBM was able to lock in DB2 maintenance fees at a set price for five years, eclipsing Oracle's two-year offer significantly and eliminating the Oracle tax for external cores. Savings on this component is estimated to be in excess of 50 percent to provide a payback for the migration effort in 24 to 30 months depending on any unforeseen expenses (and you always plan on some surprises). From the beginning, IBM has had regular, named executives personally monitoring this project and iHealth is feeling loved and cared for in a way that is a sharp departure from how Oracle/Sun was treating them.
iHealth plans to bring up the new DB2 environments in parallel (they have 68 large databases), but if the initial migrations go well and the line managers allow it, it hopes to shift to clean cutovers where it can simply discontinue the old systems. There has been little need for general retraining (users won't see any difference when this is done and the IBM hardware is easy to learn); however, the company estimates that it will take three months to transition the Oracle specialists to DB2 and, so far, they all have agreed to retrain. iHealth did hire one DB2 specialist who previously worked for Coke and it has a contingency plan to hire another specialist. IBM has promised to provide specialists during the transition if the need arises to ensure a seamless a transition.
The migration is expected to be complete before the end of the year, with initial targets for completion in August of this year.
IBM continues to roll out an aggressive plan to displace as many Sun and Oracle accounts as possible in 2011. There are a number of compelling initial results that include massive savings, better executive care, a more assured roadmap and fewer vendor taxes. However, iHealth's deployment clearly isn't done yet and several of these benefits are yet to be confirmed as a result.
One thing most of us know is that it is tough to get an enterprise advocate and, not surprisingly, IBM gets more than most. That probably says a lot about how IBM takes care of its customers. We'll try to revisit this in a few months to see how things turned out. So far, the only thing iHealth would do differently is add a bit more time to the expected completion date because a bigger buffer is always a better idea.