“Accuracy. The biggest concern for us is accuracy. We want to make sure we’re manipulating our members’ accounts in an accurate way,” said Eric Gauthier, an IT systems administrator for BECU. That stringent demand for quality control coupled with continuous tight deadlines – credits from the federal reserve, for example, have to post before the start of business every day – meant a lot of jerry-rigging, redundancy, complex systems and a little bit of luck were part of the daily routine. For some jobs, 9 workstations were individually manned and monitored, requiring an orchestration that tied up resources and introduced more and more room for human error.

More time was spent fighting fires than innovating new ways to improve the business, a situation Gauthier had faced before.

“Since I’ve been with BECU, we’ve always had a culture that says we’re always there to support the member,” he said. “I worked with places where the business picks everything and IT is just stuck supporting.”

So members of BECU’s operations staff worked with other departments – security, business – to plan a counter-strike on the thorny batch processing issues the company was facing. They found one of the big challenges was coordination.

“We did some digging and we found that people would set up Windows-scheduled tasks all over the place on different servers, and we didn’t have any visibility,” he said. Things were even worse if it was an outside group requesting a report: Requests were often sent via e-mail, passed along informally until some precious downtime was found and then e-mailed back. Human lag was a real issue.

So the operations staff, along with liaisons from other groups, settled on UC4′s ONE Automation product, which helped drill down into which servers were touching which data and gave Eric and the rest of the staff a “single-pane view” to help both optimize processes while ensuring that when emergencies did crop up, the team knew exactly where to start working.

They also began automating the routine reports: Now, instead of an e-mail, common reports could be generated and sent automatically.

Not only could the operations staff focus on more interesting work, but management was happy, too.

“They want us to automate and be smarter about the day-to-day things that can be automated, so that we can spend time researching and doing projects that give us a competitive advantage,” Gauthier said. “Things that help our membership are our sole reason for being.”

Michael Morisy is the editorial director for ITKnowledgeExchange. He can be followed on Twitter or you can reach him at Michael@ITKnowledgeExchange.com.

This method of managing provisioning and configuration involves codifying something that you’re currently doing manually. This will require different tools and levels of management as your infrastructure evolves.

Automation by Phase

• Server consolidation: Start off simple, utilizing scripts you’ve probably compiled on your own to simplify the manual processes. At this stage, vendor tools will be deployed to share these shortcuts across the enterprise.
• Infrastructure optimization: Support begins to expand beyond single vendors, and capacity management becomes an issue. Automation takes on the responsibility of patching and managing configuration.
• Automation & Orchestration: Another level of maturity in your virtualization strategy means adding provisioning, configuration standardization, compliance, running processes, application deployment, and self-service (or private cloud-like services) to the automated mix.
• Dynamic Data Center: Whether or not this means deploying a private cloud, reaching this level of maturity in your strategy heightens your application portability and calls for SLA-accompanied automation.

Not so fast

Bottom line is, if your application is faulty in the traditional model, it will be faulty in the young and hip model. There are opportunities to ease yourself into the water, however, to make sure that the applications you’re choosing to deploy atop a new infrastructure are the right ones. Rackspace’s recent OpenStack – their open source offering of their very own code – allows you to integrate the code you’d be utilizing in the cloud in your very own infrastructure. Not only can you get comfortable in your comfort zone, running the code in front of the eyes of your own team members before sending your babies out into the cloud, but you can monitor how your applications will run in that environment.

As with all technology that promises to deliver ease-of-use, there are many ways automation can go wrong. Before implementation, ensure each instance has been tested and that there are periodic checks in place once deployed.

Though the conversation at events such as Interop is dominated by cloud, there are specific and real-time practices that you can begin implementing today to prepare yourself for the somewhat cloudy future.

Melanie Yarbrough is the assistant community editor at ITKnowledgeExchange.com. Follow her on Twitter or send her an email at Melanie@ITKnowledgeExchange.com.