Big Blue unveiled Tivoli Workload and Capacity Management today. It's a self-managing autonomic software tool that automatically allocates server resources based on the ebb and flow of computing demand.
The latest management tool from IBM's Tivoli unit, one part of the company's on-demand computing strategy, allows servers to run during outages. The outages may be planned, as in the case of application updates across the grid, or unplanned, such as a natural disaster or computer failure.
Software packages from rival vendors like HP (Quote, Chart), Computer Associates (Quote, Chart) or Sun Microsystems (Quote, Chart) also react in real-time. But IBM's new application has a twist: It lets computers shift workloads to make sure the most important computer processes or transactions are seen to first.
For instance, a customer can limit a server to 95 percent capacity. When a server approaches that threshold, the software will shift the workload to another server, said Michael Morneault, global market manager for workload management at IBM.
Morneault said the software also helps customers forecast workloads for high-priority projects, helping companies plan how they want to distribute their resources. Normally, such tasks are done manually by IT staffs and may take days, weeks, or even months to complete.
The new tool is optimized to run grids based on IBM technology, but Morneault said customers who use competing products such as Sun's Sun Grid technology may also use Tivoli.
Tivoli Workload and Capacity Management is a combination of two existing Tivoli tools, Workload Scheduler and Intelligent Orchestrator. Scheduler provides the console for automating the monitoring and management of workloads in mainframe and distributed environments; Orchestrator adds the self-managing autonomic technology to deliver capacity as needed.
The new solution, which starts at $2,900, is not a new concept. It takes tips from a traditional programming technique, called batch computing. Batching, which originated in the 1950s and is still used today, calls for computing jobs to be submitted to queues and scheduled for processing.