Last Tuesday I met with Yvonne Perkins, VP of Enterprise Platform Software, IBM. We were in a one to one at Impact 2008, discussing CICS, its future and how IBM positions it and other core middleware technologies such as WebSphereMQ.
Yvonne made the quite reasonable argument that CICS was very much part of the overall IBM SOA story, and that IBM was investing accordingly; that is, heavily. SOA helps customers to extend existing investments and IBM is investing in all the surrounding tooling to make this easier.
My problem with this argument as it relates to CICS is largely down to the fact that without skills growth a platform can only ever be a cash cow. The law of leaky abstractions is a major problem for service enablement. As Joel says:
“The law of leaky abstractions means that whenever somebody comes up with a wizzy new code-generation tool that is supposed to make us all ever-so-efficient, you hear a lot of people saying “learn how to do it manually first, then use the wizzy tool to save time.” Code generation tools which pretend to abstract out something, like all abstractions, leak, and the only way to deal with the leaks competently is to learn about how the abstractions work and what they are abstracting.”
What happens if you want to change the underlying enterprise data model, for example? You can’t do that without changing the code. You can service-enable all you want, but SOA is as much about component and service isolation, enabling flexibility and portfolio maintainability, than service reuse. Martin Nally Rational CTO (As seen on RedMonkTV here) made a similar point in our unconference last week.
When I wrote a blog a while back called Mainframe = Youth, which gathered together a number of news stories that all talked to mainframe skills rejuvenation, it was clear that it was the skills revival, not the product, that was the story, as least as far as the wider market conversation was concerned.
If SOA is veneerware it is far less valuable than the hype would suggest. But how do you get under the veneer? I know that some Impact attendees were unhappy last year that Impact was too much hand-waving and not enough hardcore technical learning and advisory.
I spent some talking to customers last week, and grassroots IBMers, developers and so on, about CICS. One impassioned defence of IBM strategy came from Chris Hodgins, a young developer at Hursley, who popped up on twitter to challenge me.
“I see SOA as another opportunity for customers to take advantage of their existing assets. Rather than a move away from what we do, I see it as an embracement of current technology and using that to grab more value from the time and effort that customers have already invested into CICS. One thing that the CICS director mentioned numerous times during the CICS Q&A was that we were looking at non-disruptive technologies as a way to do just this. Hence all the focus around SOA, Atom feeds, CICS WEB interfaces and such like. There is a reason we are seeing some of our fastest adoption rates for the new release.
As I said previously, I may just be biased and I’ve only been at this a little under 3 years but that would be my take on it. I wouldn’t assume that because we are doing SOA, that we aren’t working our butts off improving, innovating and building on the traditional aspects of CICS TS 3.2 as well.
Good job Chris. I look forward to meeting you some time.
I still think IBM needs to keep banging the new workloads drum, to the mainframe middleware sub system as well as the box itself rather than just talk to extending existing investments. But Chris is proof there are young hardcore CICS folks coming along the pike.
After a couple of days at IMPACT it also became abundantly clear that the education tracks were plenty specific enough when it came to IBM middleware. The “connectivity” track was definitely solid. The balance of SOA hand-waving to technical education seemed right on the money.
I don’t agree with the argument its all about leveraging existing infrastructure, which reminds me of IBM’s to-my-mind somewhat defeatist mid 90s argument that mainframe MIPS workloads were increasing, even if z wasn’t winning new customers. IBM was reasonably happy with the status quo. The arrival of China as a major purchaser changed the terms of the debate and suddenly the mainframe had a new reason for being. Chinese National Railways has over 8m employees. A large bank is going to have well upwards of 200m customers…
One of the most interesting conversations at Impact happened over Red Bull and vodka, when I talked to Bob Madey. If I could find a decent recent bio online for him I’d point to it. I know Bob from Tivoli days. He is now in IBM App Integration and Middleware. He seems to have the bones of an evil plan for CICS, which will involve going really deep on metadata analysis and indexing, understanding the context of every transaction, not just the transaction value itself. Why did this intrigue me? It gets back to one of my thoughts about CICS. Originally it was a customer information control system. I think its time to make CICS a front and center information management platform again, rather than just providing transaction management services. That’s a potential new frame for the frame. And yes – IBM’s moves to offer REST interfaces to CICS is a good move in that direction, as is all the WS-* service enablement.
Leveraging existing investments is fine- but i guess it depends what you mean by leverage…