Autonomic decisions may not be automatic

As vice president in charge of IBM's initiative for Autonomic Computing, Alan Ganek has the tough job of convincing budget-constrained corporate IT shops to spend money on cutting-edge technologies that have yet to establish long track records of success. Even though many of these autonomic products are innately designed to save them the money, it remains a tough task.

Ganek recently sat down with InfoWorld's Ed Scannell to discuss a range of topics, including the acquisition of Think Dynamics, how he plans to convince CXOs to spend money to save money through autonomic technologies, and the advantages of IBM's On Demand initiative over competing utility computing visions from archrivals Hewlett-Packard and Sun Microsystems.

Q: InfoWorld: What is behind IBM's decision to buy Think Dynamics?

Ganek: They have built something that fits in exactly with our concept of what Autonomic Computing is about. It has the attributes of monitoring (and) analyzing and the sort of control loop we use as a basic architectural model. Their whole scheme directly applies to that. Also, it is rare to find an implementation so strongly rooted in standards. It gives us a very comfortable feeling that we can extend it to the environments that we are interested in and do so very quickly. It complements existing configuration products we have in Tivoli as well as some of the capabilities we have in our servers to allow for dynamic provisioning of capabilities. That is becoming important for what we see as increasingly spiky workloads. This is something that didn’t happen five or 10 years ago. It happened only since the Internet has been applied seriously for business. So it is only a three- to five-year-old problem.

Q: What role will they play in your overall strategy from a technology standpoint?

Ganek: We have been emphasizing recently the need to have a policy-based orchestration that coordinates various events. In other words, you can't separate the ability to drive highly available solutions from the ability to do dynamic provisioning. One of the reasons you might dynamically reconfigure and provision another system would be in response to a failure occurring someplace in order to keep that process going. So this notion of policy-based orchestration is a very broad one driven by business requirements and policies. The acquisition of Think Dynamics is a key step in delivering on that approach. They offer a standards-based implementations based on the J2EE concepts written in Java and that ascribe to Web services standards and that provide a mechanism to define the resources in the system. So it is a great fit with what we had already.

Q Don't you have the problem that many CXOs still have the dot-com crash fresh in their minds and are reluctant to spend money -- even on e-businesses that have the potential to save them money in the long run? They don't want to bite unless they see ROI in six months.

Ganek: That is why the On Demand concept is a much broader one than the Utility Computing one. We are not talking about ripping out a user's whole environment and replacing it with some monster thing. These CXOs do not need to be reminded that they are spending more money on people than on technology and the trend is continuing in that direction. That was not true 10 years ago. Ten years ago most of the costs in the datacenter were with the software and hardware licensees. Now your costs are mostly with people. There are three value propositions with autonomic computing: return on investments, quality of service availability, and time to value -- i.e. how quickly you can produce things. When we talk to users we address it in two ways. One is in the productivity of your people -- I don’t know of anyone who doesn't want to make their people more productive -- and the other, which grabs too much of the limelight, is making your technology more productive.

Q: Do many of the Fortune 100 companies get this? Are they at least thinking about moving toward implementations soon, or are most pushing it off for another year or more?

Ganek: I think people are very receptive to what they need to do in order to make their environments more efficient. Clearly, the overall economic situation is affecting everyone's spending. But I have spoken to CIO groups here and abroad and I get a strong feeling that these are important goals to them and that they are very interested in exploring how to best position their systems. Our approach has not been to paint a 10-year-away picture, but rather to focus on getting enhancements that can be incrementally applied in many products across the spectrum.

Q: Some corporate users look at these on demand or utility computing initiatives from an IBM, Sun, or HP as just another way for vendors to get them to spend money on new equipment and software or to drum up business for their services groups. Is that a reasonable claim?

Ganek: For IBM's part, I don’t believe so. I do hear that until I sit down with them and they compare us to other things being offered. There are offerings out there that are exactly as you describe, approaches that require a green field environment where you must start with racks of systems that are within certain specifications and/or build a new environment using proprietary software. And that is a big bullet to bite. But we have not taken that approach from the outset.

Also, it is important not to equate utility computing with on demand. It is one of the vehicles that users can pick if that meets their requirements. The approach we have taken is to recognize that the complexity we are dealing with is in all the components, then build a framework for making these components more self-managing. We attempt to enhance not only the systems management tools but the components themselves. We are focusing on both management and manageability so you can put the components together in a way you want. We do have a set of services offerings, but nothing that we have talked about in terms of autonomic products requires you to use them.

Q: Because the concepts of on demand and utility computing are new and users have not had the time or inclination to examine them closely, is the problem mainly one of marketing?

Ganek: There is always the challenge in this technology industry to tell your story and avoid the confusion with other stories being told. We have been working hard to educate the press and analysts so they understand these things and working with our users to give them proper perspective. In fact, that is why we have leveraged customer councils to define a lot of the technology we are putting together. They tell us where the real problems are and what their priorities are. We are finding no surprise in the wide diversity of their needs. There is not a one-size-fits-all world out there, and that is why you need an approach that is flexible. I think (IBM Chairman) Sam Palmisano was careful to define (on demand) as a state of business flexibility that allowed for business spontaneity and the ability to be resilient and react to different situations. It did not prescribe any one technical solution.

Q: Are there any concerted efforts among IBM, HP, and Sun to work together to ensure that some of your competitive offerings in this space will work together?

Ganek: Those efforts are conducted through standards bodies for the most part, and there is active participation among the major companies on a number of key standards. One, for instance, is Web services, where we have collaborated heavily with Microsoft. The Global Grid Forum also has participation from the major vendors, and the Distributed Management Task Force is one other example. One of the key standards for heterogeneous workload management is called ARM, or the Application Response Measurement. We collaborate with HP and others on that. And then of course there is the whole J2EE platform.

Q: The problem in working through standards bodies is all the foot-dragging and political infighting that takes place, which is why it takes a long time to get anything done. How can you get around that in this case?

Ganek: The way we approach standards is that code talks. So you get to the standards bodies but you push implementations that are meaningful out into the marketplace. That is what gets standards completed. You can put people in a room to talk standards where there is no market pressure, no customer deployments, and so you are right that people can argue forever on some issues. But when you have compelling offerings and you are bringing them to the marketplace and customers are excited about them, then you can move these things through quickly.

Q: How does Microsoft fit into the mix regarding on demand or utility computing? They are not total solution providers, like IBM and HP, that can deliver software, hardware, and services. How focused is IBM on them as a competitor?

Ganek: Microsoft is certainly an important force in the industry and we compete with them and partner with them. I don’t expect that to change. Clearly, you expect their focus to be on the client and with the client attached to departmental servers and so forth. I think the collaboration with them on Web services is an important linchpin in terms of tying things they are doing in with what others are doing. I certainly welcome their focus on issues that relate to self management as it relates to the client because we can do a much better job if they are supporting that as well.

Q: How important will Web services be as a factor going forward for integrating these competing initiatives?

Ganek: It has a lot of momentum and can be a very powerful lever for achieving some of the goals we are talking about. The ability to clearly detail how you define a service, how you locate it, how you attach to it is ideal for building these environments where the resources have a virtualized relationship to each other. You can then adapt to them and plug in various engines for analysis control, configuration, and provisioning without having to disrupt every underlying component. It is a very powerful thing. There have been many efforts over many years on distributed services, but this one seems to have the most weight behind it and is built on solid approaches. It seems to have a strong buy-in from enough key companies that it gives me a very optimistic view of its ongoing importance.

Q: What sort of mode is IBM in right now in terms of buying or building what it needs to fill in the pieces of its on demand strategy?

Ganek: It is hard to answer that question generally. We always look at the buy vs. build question from a pragmatic and business standpoint. We have a talented team with a broad set of skills, but it would be foolish to think you can build everything yourself because there are time-to-market issues. We routinely (look at) what is out there and try to make judicious choices, as was the case with large acquisitions like Rational or a small one like Think Dynamics.

Join the newsletter!


Sign up to gain exclusive access to email subscriptions, event invitations, competitions, giveaways, and much more.

Membership is free, and your security and privacy remain protected. View our privacy policy before signing up.

Error: Please check your email address.

More about Distributed Management Task ForceEdge TechnologiesGlobal Grid ForumHewlett-Packard AustraliaIBM AustraliaMicrosoftProvisionSun MicrosystemsTivoli

Show Comments