Summary:

Cloud computing hopes to deliver convenient, on-demand access to shared pools of computing resources that can be provisioned with minimal effort. But since the virtualized environment becomes even more dynamic, complex and real-time in the cloud, the benefits of IT analytics become even more dramatic.

cloudpuzzle

Edit Note: This is the third of a three-part post. In Sanna’s second post, he described a new virtualization management approach based on advanced IT analytic applications and powered by behavior learning technology. Big IT organizations are using it, but does this apply to the cloud where things become even more complex?

Cloud computing hopes to deliver convenient, on-demand access to shared pools of computing resources that can be provisioned with minimal effort. In fact, some solutions envision a “self service” interface where users can drag & drop virtual data center service components in a build-your-own-cloud environment. Since virtualization underpins most cloud solutions, the role and benefits of behavior learning and predictive analytics as the basis for performance management in the cloud are the same. But since the virtualized environment becomes even more dynamic, complex and real-time in the cloud, the benefits of IT analytics become even more dramatic.

Private cloud-based infrastructures are far more complex and difficult to manage than anything a data center manager could have envisioned even five years ago. While the cloud may simplify service provisioning, the underlying infrastructure has many more interrelated, moving parts to oversee:

  • Applications may be provisioned in virtual server clusters that can dynamically grow in number or change in available computing power during peak usage times.
  • Physical servers can have widely varying workload profiles as virtual machines are created and shut down, or their workloads are moved from one host to another.
  • Network and storage infrastructure will be stressed, as virtual server activity generates tremendous I/O traffic between physical hosts and storage infrastructure.
  • Diagnosing customer experience and application performance will be more difficult due to the complex, dynamic relationships between applications and infrastructure –both physical and virtual.

In such an environment, with optimized resources, there will be little room for error. In such a complex environment, it is inconceivable that you would continue to try to monitor performance of the IT infrastructure and applications via manual rules-based processes. That’s why a performance management platform based on behavior learning technology and leveraging advanced statistical analysis and predictive analytics become necessary. It can:

  • Eliminate labor-intensive management tasks by self-learning and adapting to the changing operating behavior of each component – physical or virtual;
  • Detect behavior anomalies and send alerts to impending problems before services are affected;
  • Automatically correlate the behavior of components in heterogeneous environments and provide a means to automate the virtual resource provisioning and right-sizing of infrastructure.

The benefits of this last feature cannot be underestimated when deploying a virtualized “cloud” infrastructure. Automatic resource provisioning to maintain negotiated service level agreements requires accurate analysis and data. Today, this analysis is mostly manual and cloud management teams are spending most of their time studying and analyzing their environments before “pulling the trigger” leaving little time to architectural improvements. The promise of a mathematically proven method of automation liberates resources to perform more productive tasks.

Also, many enterprise IT leaders need to recognize that to realize the full economic benefits of their virtualization projects or the promise of the cloud, they must find a new way of performance management that is not manual and rules-based. Every day they hear stories about organizations that attempted to virtualize and simply traded their physical sprawl for a virtual version of the same environment.

Because of these new, analytics-based performance management solutions that will come of age in 2011, CIOs will be able to serve their customers (line of business owners) more efficiently. CFOs will be happy with lower hardware costs and energy bills from “right-sized” infrastructures. And application owners will be able to confidently deploy or change resources in minutes, not weeks. Private and public clouds are here. Now, the new breed of analytics-based management frameworks will enable IT to keep promises and show off the true benefits of cloud-based computing.

This is final post in a three-part series. The first post ran here, and the second one can be found here.

Nicola Sanna is chief executive officer of Netuitive, and has held the role since 2002. Netuitive enables enterprises to proactively manage the performance and capacity of their IT infrastructures – physical, virtual and cloud.

Image courtesy of Flickr user Adam_T4.

Related content from GigaOM Pro (sub req’d):

Comments have been disabled for this post