Evangelizing Mainframe
Print Email

The Future is Hybrid Computing

Does your organization keep metrics on productivity? How about the time to troubleshoot a problem and the expense of the people involved in solving complex problems? If not, maybe it’s time to examine what you are spending on the management and troubleshooting of your applications and infrastructure.

The logical question is, why? I have had several customers break their IT budgets down to application maintenance and implementation of new function. In most cases, the vast majority of time and funds are spent on maintenance, and little is spent on new function. In some cases I’ve seen as much as 85 to 90 percent of the entire IT budget spent on maintenance.

The real problem here is figuring out how to free up more of the budget to serve the business and help it grow. I attribute high maintenance costs partly to infrastructure complexity. Having personally experienced many system outages in recent days, it’s easy to believe there’s a better way.

The more an organization decentralizes computing, the more infrastructure it takes to run applications, and more infrastructure means more people. Even with the maturity of virtualization technologies, I routinely run into organizations with thousands of servers. Teams are segregated into specialty areas such as security, network, performance, OS, and on and on. Typically, when an ambiguous system problem is causing a major business issue, a cross-functional team is put together to troubleshoot and solve the problem.

In many cases, each team has specific tools it prefers as best of breed for troubleshooting a specific area, but those tools don’t correlate to any of the other areas. Because teams are organized in silos, the natural instinct for each technology area is to prove the problem wasn’t caused by their infrastructure, rather than collaborating across the teams to solve the problem. This process can lead to days, sometimes weeks, of back and forth before someone stumbles upon the real problem. Then you have the root-cause analysis, which can take even longer.

So why not take an end-to-end application or set of business functions, put them in a single container, and use a single set of monitoring and management tools to run the infrastructure? It sounds too good to be true, right? Wrong!

If you’ve been paying attention, you’ve seen the recent IBM announcement regarding PureSystems, if you haven’t, crawl out from under that rock and take a look at this. In a nutshell, IBM PureSystems are a prepackaged set of x86 and Power servers, storage, networking and middleware delivered with a single set of management tools.

You may also remember the IBM zEnterprise System announcement from 2010, which takes this concept even further by packaging a complete end-to-end IT ecosystem in a single system complex. The zEnterprise package includes a virtual desktop infrastructure, x86 and Power blade servers, appliances, virtualized Linux and Windows, as well as core business systems running on z/OS. This zEnterprise ecosystem comes pre-packed with monitoring and management tools to greatly simplify the infrastructure and labor it takes to support that infrastructure.

This is hybrid computing. It’s recentralizing the IT infrastructure. It’s reducing complexity. It’s allowing organizations to focus on new functionality vs. never-ending maintenance and troubleshooting. It’s saving money. And it’s here now!

IBM has solidly addressed the concerns of customers and provided truly enterprise-ready hybrid computing models aimed at realty reducing complexity and infrastructure costs. Organizations can now funnel more of their budgets to business development and growth rather than throwing capital at a never-ending server and infrastructure sprawl.

Chris Gombola is Enterprise IT Architect with 13 years of experience at IBM.

Posted: 4/24/2012 3:51:04 AM by Chris Gombola | with 1 comments

Print Email

Please sign in to comment.

Sign In


Comments
Thanks Chris - nice summary of hybrid computing and you opened the discussion with a great point on metrics for troubleshooting complex problems.
Posted: 4/26/2012 6:27:43 AM by Kathy Bazinet