IBM Executes On On-Demand

"It's a roadmap of how you take an idea and turn it into a set of products and services that people want to buy," says Ted Schadler, principal analyst at Forrester Research.

Last October, IBM chairman Sam Palmisano first uttered the phrase, "e-business on-demand" in a widely reported speech to a customer group. Since that time, the "on-demand" concept has infiltrated the industry lexicon, due in large part to the colossal media blitz IBM has put forth.

But truth be told, many solution providers are still trying to figure out what this fuzzy concept of on-demand really means for them in terms of products and services. Today, IBM stepped up with some answers. Big Blue officials explain e-business on-demand as the optimal IT infrastructure, a system where networks and software serve business needs through automation, self-monitoring, exploiting unused capacity and operating in real time -- much like a living organism.

News announcements around storage, blade servers and applications infrastructure reflect the IBM refrains around application and network integration, Web services, autonomic (or self-monitoring and healing) systems, and grid and utility-based computing.

Among them, the company took the wraps off IBM Adaptive Server Allocation for WebSphere Application Server, which embeds grid technology into WebSphere to help maximize the use of servers more efficiently by drawing on them as a unified pool of resources. It's due out later this quarter.

In explaining how the server allocation tool works, IBM outlined its work with financial services giant Schwab. At Schwab, they are distributing the workload of its wealth-management system from one application server across to four application servers. With the systems working in parallel, the company has been able to reduce the applications processing time by 94 percent, according to Bob Sutor, director of WebSphere infrastructure software at IBM.

"This offering is a way of dealing with your application servers like a grid, so that the workload is distributed dynamically across systems to make sure that you are meeting your SLAs," Sutor says. "If you can do this, you can increase the utilization of existing Web servers, which may mean you can use fewer of them."

Capabilities from the company's Tivoli group have been built into WebSphere to help provide workload balancing and proactive monitoring of server availability, usage patterns and priorities.

On the storage front, IBM executives unveiled the IBM TotalStorage Virtualization family of products, indicating that its much-talked about and long-awaited storage virtualization product -- code-named Tank -- may finally see the light of day. They announced two software pieces of the family: the Total Storage SAN Volume Controller, which is scheduled for release this June or July, and the Total Storage SAN File System, scheduled for release this coming December.

"For the first time in 60 years, we are breaking the umbilical cord between the server and storage," says Jens Tiedemann, IBM's vice president of storage software.

The basic concept behind virtualization is about breaking the physical boundaries within storage that make a storage administrator's job a grueling task. Servers have such a blind-sided view of its own storage -- whether it is direct-attached or SAN-attached -- that storage administrators must go through the tedious task of bringing down a server, and the application running on it, in order to do a simple job of expanding capacity.

"There is such a huge management task in handling 1,000 servers and which storage is allocated to it," Tiedemann says. "This is really what makes a storage administrator's job a 2 a.m. job instead of a 2 p.m. job...Some disruptive technology needs to take care of that."

Management software has become a key market. Consequently, IBM's announcement is one of many on the industry today, since both new and old storage vendors are jumping into the storage management software market. Big Blue's SAN Volume Controller, which works across disparate disk arrays, automates the process of keeping the application running when basic jobs are in progress such as formatting a new disk, reallocating storage and making remote copies of data.

The IBM SAN File System, designed for heterogeneous storage networks, helps create SAN-wide file system.

IBM executives say that thus far, at least 30 partners already have SAN Volume Control software installed in their IBM Total Storage Solution Centers, which function as interoperability labs for partners and customers. By June or July of this year, the company expects that figure to increase to more than 50.

Other announcements include the IBM Standby Capacity On Demand offerings for blade servers and storage, which enables customers to buy fully configured blade or storage systems at a reduced price, then to turn on additional capacity over a six-month period as needed. This offering is due out later this year. IBM also unveiled the Web Server Provisioning automation tools that leverage autonomic functionality to help customers better manage servers.

All told, analyst Schadler lauded IBM for coalescing its entire company around executing the on-demand vision, from the software and systems groups to IBM Research and Global Services. And, he says, IBM makes a more credible case for its vision by eating its own dog food.

"It's important that they are using these principles internally, in their own business," Schadler says. "And they are cutting costs out of their huge supply chain and other operational areas."