Top 2009 Stories: Exits and Migrations
December 29, 2009
We're practicing the newsman's tradition of reviewing 2009 stories to select those most important for the year -- as well as the years to come. At week's end we'll look at our '09 forecast to see how well we predicted your year, as well as choose the most significant stories of this decade.
To start an '09 review, let's consider two items important to the future of the 3000 Transition: moving onward as well as away from the platform, both from an HP position as well as the view of the independent community.
HP exits lab business, leaves source license promises — in January the doors on the development lab for 3000 work got shut at HP, even though the lab itself doesn't have doors, and the engineers in cubicles remained employed at HP doing other work. But 2009 was the first year the vendor wouldn't deliver patches for the 3000, software development HP performed in its Cupertino offices.
At the same time that HP ended its creation of the MPE/iX heart of the 3000, the company outlined its conditions and restrictions to license source code for the operating system to third parties. A key software tool, SS_UPDATE, is not part of the license plan. HP's best offer was read-only source, to be used for support purposes and not for creating new MPE/iX features. The license terms were secret, negotiations were covered by confidential disclosures, and HP got unhappy when the OpenMPE advocacy group announced it was halfway to raising its money for the source code license fee.
A few weeks later, HP's Jennie Hou announced that the vendor will announce the winners of the licenses by the end of March, 2010. Support companies and software icons of the community will wait and see what difference seeing the MPE source might make, but source should simplify workarounds for companies both homesteading and doing long-term migrations.
Migrations were extended and restarted, as well as simplified — A colossal crash of a $14 million migration project at Washington State colleges came to light during the year, with HP's oversight leading a series of misjudgments from third parties struggling to move antique Transact and Protos source code. Oracle and .NET were the targets of the plan started in 2003 and declared dead by the end of 2008. The failed project illustrated the risks of attempting too much change. Word is emerging about a restart of the migration, with work led by different outsourced providers.
The most frequently successful migrations involved a lift-and-shift mentality, in terms of the numbers of customers moved away. Speedware talked about a project to get ING Australia onto HP-UX, with a report that the customer insisted even the bugs in the existing application be moved to keep auditors at ease.
ScreenJet's Alan Yeo echoed that experience at the e3000 Community Meet, saying that a migration poses less risk when it changes very little functionality. "We learned that the the bigger a company gets, the less they can afford to change anything," he said. "The cost of retraining staff, or changing the way a business function works, costs them more than doing the work of the migration,” Customers saw that re-engineering is Phase Two of a migration.
Speedware estimated at the Community Meet that it sees only about 1,000 customers left on the 3000 platform, based on its research into its customer and prospect lists. Other community solution providers felt the numbers could run up to 5,000 customers remaining. HP continued to be surprised that 9x7 customers were coming out of the woodwork. The Afterlife of HP's 3000 business carried on into an eighth year.