lustratus LiteBytes

What a surprise….HP, Oracle and Intel…

…are targeting legacy (ie mainframes). According to the news article on SearchWebServices.com, the threesome have built an SOA reference architecture to help consultants convince CIOs to

transition from COBOL programs running on mainframes to a service-oriented architecture.

Indeed, HP is quoted in the following paragraph,

John Pickett, worldwide manager for mainframe alternative programs at HP, said the model based on the reference architecture gives IT executives a demonstration of what SOA could do for them. It also shows the viability of SOA as a replacement for tried-and-true legacy mainframe systems, he said.

In its press release, HP is a little more circumspect in its language, pointing out that the new Application Modernization Initiative uses SOA principles without actually stating that this is an initiative to ‘transition from COBOL programs running on mainframes to a service-oriented architecture’.

And yet, I suspect that this is indeed the intent. I think this HP/Oracle/Intel initiative is trying to spin the SOA message to imply that you are either using COBOL mainframe applications, or you are operating in a SOA environment. In other words, SOA means ditching your mainframe COBOL apps. Admittedly, there is some admission that this may not be the best in the near term for some apps, but the implication is still there.

If my suppositions are correct, this is yet another attempt to twist the SOA message into a self-serving piece of marketing. In fact, companies can quite happily be operating an SOA where COBOL mainframe applications play a full and active role. The free Illuminatus Research report on mainframe SOA tools discusses this scenario in detail. The whole point of SOA is to offer an environment where services can run wherever it makes most sense, preferably as non-invasively as possible (ie leave it alone if you don’t have to touch it). This does not mean move everything off the mainframe.

What really irritates me is that when people talk of delivering improved performance, lower cost of ownership and better flexibility by moving applications off the mainframe, no-one ever seems to bring up the thorny question of RISK. I plan a separate blog entry on this.

Steve