A Cloud With a Brain
CloudBurst 2.1 can help move IT from cost center to revenue source
Illustration by Tony Gold
Although IT departments play a crucial role in nearly every aspect of revenue-generating operations—from processing transactions and generating business intelligence to developing, customizing and maintaining software—all too often they’re viewed as cost centers. The time has come to change that. IBM CloudBurst* 2.1 jump-starts IT departments into the world of cloud computing, enabling them to flexibly allocate resources in minutes, maximize utilization, minimize system-management time and speed the deployment of new revenue-generating operations.
Released in December 2010, CloudBurst 2.1 for the Power Systems* platform offers enterprises everything they need to establish a cloud-computing environment powered by a processor that redefines performance. It’s a scalable, integrated solution with preinstalled, preconfigured hardware, software, networking and storage, plus IBM QuickStart services to help remove the guesswork from migrating to cloud. CloudBurst simplifies the move to a virtualized system, helping IT departments drive down costs while freeing them to focus on delivering next-generation business services.
Work Smarter, Not Harder
On average, 70 percent of every IT dollar goes to managing the current infrastructure rather than investing in new hardware to increase capabilities. Of course, new hardware doesn’t necessarily bring as much benefit as executives might expect. Studies show that in a distributed computing environment, as much as 85 percent of capacity may sit idle at any time, generally as a result of overprovisioning to accommodate workload spikes. Companies also tend to buy more capacity than needed in an attempt to future-proof their systems. In a single server, low utilization rates cost money. Replicate those issues over a hundred or so boxes and you’ve got a serious problem, not just in terms of wasted capital investment but also in increased operations cost—the machines gobble up power, space, cooling and maintenance.
Cloud computing provides a solution. The approach leverages virtualization and networking to centralize computing capacity and resources. Consider software test-and-development organizations, which typically consume 30 to 50 percent of IT resources. In the old model, an internal group supporting SAP applications, for example, might request its own machine to maintain and upgrade the software. In a cloud model, all computing capacity belongs to the cloud. Instead of buying the test-and-development group its own server, capacity and software licenses, the IT department can just provision a virtual machine from the capacity pool. If demand rises, it’s easy to give the virtual machine more processing power or memory because it’s only a part of the larger system.
With cloud computing, IT departments can drive utilizations as high as 90 percent and cut in half the labor for configuration, operations and management. The cost of IT support to the end user can drop as much as 40 percent. “The end user gets better service and the data center reduces its cost of delivering the services,” says Steve Curtis, manager of worldwide marketing for virtualization and service delivery at IBM. “It’s a win-win story all around.”
Particularly when coupled with data-center consolidation, cloud computing provides big benefits. POWER7* technology, for example, delivers three to four times the performance as POWER6* technology, in the same footprint and energy envelope. With the POWER7 platform, enterprises can put workloads from numerous physical boxes onto one server partitioned into multiple distinct virtual machines with a variety of operating systems. Designed from the silicon up for virtualization, POWER7 systems provide unique workload optimization, 99.997-percent availability and enterprise-level security, making them the ideal platform for cloud computing. “You might have a hundred distributed systems you put on 10 POWER7 boxes,” Curtis says. “As you consolidate, you use less energy because you’re using fewer systems. You have fewer cores so you save money on software licensing. You need fewer administrators because you don’t have as many individual boxes to administer. But because of virtualization, you have even more flexibility than offered within traditional IT environments. POWER7 servers offer a lot of advantages to deploying a cloud-computing environment.”
That’s the theory, anyway. The reality is a bit more challenging. Building a cloud environment is nontrivial and can commandeer IT resources that would otherwise be spent on the enterprise’s core business needs. Once the cloud is in place, systems typically get provisioned manually, which requires time and effort from already overworked system administrators. Perhaps worst of all, processing requests takes time, which generates a large backlog—and frustrated internal customers. Those customers quickly learn that the process of obtaining new capacity is slow and torturous. As a result, once they get computing capacity, they tend to hold onto it even after their projects are done. This reduces the overall availability of resources systemwide, which makes getting new capacity even more difficult, and so on. The result is a Dilbert-esque situation of companies investing in new capacity even as current capacity sits idle.
Imagine a different model, one in which test-and-development users simply log on to a self-service Web portal and click through menus to order the capacity they need. Standard requests get a blanket approval. Provisioning takes place automatically, without delay. When the project concludes, the resources get returned to the pool. And all of this takes place without significant intervention by the IT staff. That’s the advantage of IBM CloudBurst.