Late last year IBM Software Group (SWG) “acquired” IBM Systems and Technology (STG) group – with software head honcho Steve Mills taking over management of the combined group. The move had been coming for a while, and the parallels with Oracle’s takeover of Sun Microsystems last year are clear. The market is showing a new tolerance, perhaps even a desire, for vertical integration up and down the entire IT stack, from chips to servers to networks to storage to databases and apps. Call it the Apple effect. The golden age of Open Systems is arguably coming to a close – with user experience (UX) now seen as more important than portability, or freedom to leave.
We’re not exactly regressing- the Internet is of course built on open standards, and systems and software integration is becoming easier than ever, using Web technologies – but the pendulum is swinging towards more proprietary approaches delivering better UX.
Great UX is a function of masterful packaging. Its not about building the best components but doing the best job of integrating them. Open Source, somewhat counter-intuitively, is a driver of the New proprietary – Apple is built on FreeBSD foundations, for example, while Google runs its own customised Linux, and so on. The cloud accelerates the new proprietary – because, not being classed as software distribution, it allows for open source code to be embedded to deliver a service, without requiring any reciprocity.
IBM is of course a past master of integrating open source into its proprietary middleware products – WebSphere Application Server embeds the Apache Web Server, for example. IBM’s Smarter Computing products don’t only embed open source however- IBM’s Cloudburst private cloud provisioning engine, for example, embeds VMware’s hypervisor.
For all IBM’s ambiguity about public cloud as enterprise ready, its good to see Tivoli’s understanding that customers will put in place multiple cloud providers, for multiple sourcing strategy, for disaster recovery and cost management. One of the best expositions I have seen of this Redundant Array of Inexpensive Clouds strategy comes from Mark Masterson. Smarter Computing is also part of the post-brand IBM. So Tivoli Service Request Manager, which is at the heart of its cloud management suite, is 95% identical to Maximo Service Desk. Other elements in the IBM Cloud suite include Tivoli Usage Monitoring. The new IBM watchword is architecture, which is shorthand for packaging.
IBM’s internal goals for Smarter Computing include a stronger focus on competitive attack – IBM wants to aggressively target Oracle and VMware. But what about business focus?
One of the contexts for Smarter Computing is Big Data – as IBM puts it, the Era of Insight for Discovery, “new insights and innovation delivering huge societal and business value,” bridging into IBM’s broader Smarter Planet narrative.
In practice that means analysing a wider range of datasets than is currently cost effective. why not mine and integrate blogs, or telco CDR switches, or smart meters, or real time financial feeds?
A related context for Smarter Computing is what IBM dubs Optimised Systems – which is about applying the best technology for a particular workload. Cloud can be a bit mainframe-like here. IBM, probably more than any other company has long understood the value of specific architectures for specific roles- thus the mainframe is great for I/O intensive workloads, while IBM’s Unix boxes are more about being CPU rich.
Smarter Computing means reapplying everything has IBM learned from 100 years of resource constrained computing in a new context. With the cloud everything old is new new again. Even IBM.
Before signing off I should point out a major concern about IBM’s decision to stop talking about products and start talking about “architectures”. This kind of narrative is fine for CIOs – but what of the practitioners that work for them? The sysops guys might have excellent skills around Tivoli Service Request Manager, but wouldn’t even know IBM was pitching the same product for cloud management. The danger of Smarter Computing and the post sub-brand IBM is that it forgets the people, the operators, the practitioners as influencers and doers. IBM may indeed have too many SKUs, but if it loses the practitioner in the process of selling packages to the business executive, then it has a problem on its hands. After all the best people to really understand the value of workload specific architectures are those very same practitioners.