Q&A: CA's CTO talks about on-demand computing

Computerworld - LAS VEGAS -- Yogesh Gupta, chief technology officer at Computer Associates International Inc. in Islandia, N.Y., talked this week at NetWorld+Interop about CA's announcement of six Unicenter management software products to give companies on-demand computing capabilities.
What is really new about these six Unicenter products, and how does CA define on-demand computing? Very simply, what we have created is a set of solutions that let people manage their existing IT infrastructures in a way to get to on-demand computing. What does on-demand mean? Really, one of the big issues that comes up is that customers have large numbers of servers that are underutilized. Some say 10% of their NT servers are used. That means there's tremendous opportunity if work can be done to use those servers better. Using them better is very hard, because if you take one NT server running Exchange and want to convert it to a financial services application, that process takes days and weeks and sometimes even months. And IT shops design for peak requirements, so if the peak requirement for one application is 20 servers, they give it 20 servers -- even though the regular requirements might be three or five servers, but every now and then the application hits a peak. What the six products work together to do is identify and keep track of which IT resources [are needed by] which business processes at any given point in time.

Yogesh Gupta , CTO at Computer Associates International Inc.

We have offered historically [the] ability to do this manually. The problem with that is that if things are changing, things can get out of sync. What we've created is an automated way of figuring that out. So we've got two products, Unicenter Network and Systems Management 3.1, and Unicenter Management for webMethods 3.0, a brand-new product.
You list a second new product, Dynamic Reconfiguration Option. What is that? It acts as a brain for all of this. It tracks what the resource utilization is, what business processes need more resources and what processes need less. It does automatic capacity monitoring and resource monitoring. As it notices that some application needs more resources, it reorganizes the underlying IT infrastructure. So it automatically says, "Aha! This guy needs another server; let's give it to him."
It handles all the components of the network? It handles allocation of servers, switching to make it happen, bandwidth requirements, any IT resource, yes. If it's a router, you might have to change what part of a router supports what part of the network. If it's a server, you might have to reinstall the OS, or new applications, or change IDs. This automatic reconfiguration figures out what needs to be done, and Software Delivery 4.0 updates the software configuration
Some customers talk about organizing large numbers of servers to charge back business units for system usage in a dynamic setting. How does CA handle that? Our new Asset Management 4.0 keeps track of which business used these servers today and these routers yesterday. So, currently AXA Group uses 100 servers to support a group today and will be able to charge by the drink with this, saying you used 20 today and 30 another day. IT groups can say, "I see you use these servers for this function so I can charge back what you use." Historically, what asset management has done is work in a static environment. But with a constantly changing environment, this works dynamically to keep track.
And you have a new service desk revision? ServicePlus Service Desk 6.0 updates automatically IT systems for human interaction on the service desk supporting all the new features in the announcement.
So what is new in the revisions generally? What is new in the revisions generally is we added automatic and dynamic features for changing IT environments.
Why not market them as one product? They are sold as separate products so customers can license one or more. We decided to sell separate products within Unicenter two years ago, and it has worked well. But all these pieces work well together.
What do people most need to understand about on-demand computing? Other vendors use other terms such as self-healing and autonomic computing. How does on-demand compare? There's a lot of rhetoric about on-demand and self-healing and autonomic computing. The key thing here is that this Unicenter product set works in the existing infrastructure and doesn't require a company to start over with new hardware or platforms. Second, it is not limited to a single platform. It is truly managing in a heterogeneous environment.
Is this software really self-healing, in the sense that if a printer goes off, there's a script at the ready to turn it on? There is a healing part to this software when something is broken. When a server fails, we try to bring it online, and if we can't, the software will automatically allocate a different server to do the work. But it's more than self-healing. It is really delivering the right service level and bringing maximum utilization. It is optimal utilization of networks and systems.
How important is this new product set going to be for CA? Well, Unicenter brings in over 40% of CA revenues and is our biggest product. So, this is the evolution of Unicenter to the next level, and security and storage features will follow.
Are you late with on-demand computing compared with other vendors? If you want to talk about the rhetoric, we're definitely late with the rhetoric, but first with real products. All these ship by June, with three available now. We are absolutely the first with products, and I don't know of anybody else who actually has products like this. We made a conscious decision we would not talk about autonomic computing until the day we had products.