Losing Applications in the Cloud: The Growing Necessity for Capacity Management

Drew Robb

Updated · Jun 28, 2012

In the good old days an enterprise application was given its own server – the more mission critical the app, the bigger the server and the greater the number of CPUs and memory. That model proved highly inefficient, however, and along came VMware.

Suddenly, physical servers could be virtualized. Instead of one app on one server, many apps could run inside a single box, each on their own virtual machine (VM). Now more and more resources are virtualized – the cloud makes it possible to access networking, storage and compute resources within a virtual pool.

 “The cloud turns IT into a black box that is responsive to policies and provides services to users,” said Gartner analyst Thomas Bittman.

While virtualization offers many benefits, it also poses some new challenges. As more and more abstraction layers are applied, control of applications grows rather fuzzy. If the server is virtualized, and the network and the storage, how much control do you really have over your ERP or CRM performance?

IT, of course, can apply the time-honored solution of throwing lots of hardware at the app. But that defeats the purpose of virtualization. Nobody wants a 10 percent utilized server “just in case” demand ramps up in six months.

Back to the Future: Capacity Management

Thanks to virtualization, a discipline from the mainframe days called capacity management is making something of a comeback. Capacity management enables IT to create a reasonable estimate of application growth patterns and predict how much hardware a given app will require and when. At its highest levels, modeling can be done by capacity planners that predict when servers will reach their limit, when users will notice application performance issues and how much RAM or storage should be added and when in order to avoid latency issues.

“If you do a good job of capacity management, you can manage your applications better,” Bittman said. “There is more need for capacity management today than ever, due to the advent of virtualization and the cloud.” 

Jean-Pierre Garbani, an analyst at Forrester Research, explained that capacity management is vital when deployment decisions are being made to ensure “right sizing” of applications occurs – that is, apps are given the right amount of hardware support, neither too much or too little compute power. Costs soar when too many resources are utilized, but performance problems dog applications with not enough power.

But capacity management goes further than that. When there is an issue, the capacity planner can use tools such as TeamQuest Analyzer to isolate the source of an application bottleneck  and avoid unnecessary and expensive upgrades. It can be quite costly to upgrade the network when all that is required is another CPU or a little more storage, for example.

Benefits of Capacity Management

Similarly, the modeling of capacity can help management determine the best options for a planned upgrade. By running multiple what-if scenarios, it is possible to highlight what will be the best path forward for organizational expansion. Several possible alternatives may come out of this: an expensive option with the highest gain in performance; a low-cost option with reasonable performance gains; and a middle ground. Armed with this data, management can then determine the pros and cons of the various possibilities and ultimately make a well-informed decision.

“In one example, a company realized that its insistence on a very high level of application responsiveness would cost $5 million more than a middle-ground solution which still offered excellent performance,” said Jerred Ruble, CEO of TeamQuest. “Management then scaled down its expectations slightly and saved significantly in the process – without impacting the end user or customer experience.” 

Forrester’s Garbani sees capacity management becoming even more important in the coming years. He thinks it can be the link between IT and lines of business – a way to translate IT actions into business value.

“Financial information must be available to drive technology choices,” he said. “In a virtual world, capacity planning becomes a necessity.”

But he calls upon capacity planners to change their tune. Those engaged in capacity management have traditionally been component driven, thinking mainly in terms of CPUs, RAM and disk. They need to become business driven.

“Virtual and cloud technologies require a holistic approach that is business-service-oriented, not component limited,” he said.

Keys to Effective Capacity Management

Garbani laid out some of the requirements for effective capacity management:

  • Placement of adequate virtual and physical resources for a given application;
  • Resolution of conflicts between virtualized applications running in the same host;
  • Integration of power, cooling and floor plans into the capacity plan;
  • Creation of service-oriented capacity plans, which will let the organization determine the best potential platform for the deployment of new and existing services.

What Garbani is really talking about is the unification of long-disparate disciplines within the organization. Storage, server and networking silos must come together. Not only that, but applications will more and more be delivered as services in the cloud, an approach that necessitates the incorporation of business process and financial information to make more informed decisions.

This is far from mere theory. One of the biggest telecom providers in the U.S., for example, requires a TeamQuest model as part of its standard procurement process. No purchase order for IT can be signed without a full workout of current and predicted growth, with that translated into costs for various levels of application performance.

“Cloud technologies give IT a choice of delivery platforms that offer a variety of capabilities and costs,” Garbani said. “As with all choices, IT needs a rationale to exercise it. Capacity management will let IT leaders make the most of their cloud investments by understanding how much or how little they need to source to deliver the quality of service expected by the business users.”

Drew Robb is a freelance writer specializing in technology and engineering. Currently living in California, he is originally from Scotland, where he received a degree in geology and geography from the University of Strathclyde. He is the author of Server Disk Management in a Windows Environment (CRC Press).

Drew Robb
Drew Robb

Drew Robb is a writer who has been writing about IT, engineering, and other topics. Originating from Scotland, he currently resides in Florida. Highly skilled in rapid prototyping innovative and reliable systems. He has been an editor and professional writer full-time for more than 20 years. He works as a freelancer at Enterprise Apps Today, CIO Insight and other IT publications. He is also an editor-in chief of an international engineering journal. He enjoys solving data problems and learning abstractions that will allow for better infrastructure.

More Posts By Drew Robb