IT managers need to grease the wheels a bit to enjoy the benefits of automation. The tools available today rely on IT managers understanding what they have in their environments, what it all does and what they need it to do on its own. Inventorying infrastructure and system components and understanding all their requirements would challenge even the most savvy IT guru.
Here are tips on how to get automation technologies in place and working for you.
1. Identify repetitive tasks
Not all IT automation tasks lend themselves to automation, but there are a lot that fit the bill perfectly.
Automation has long been used to monitor systems and run batch jobs, but today’s automation expands the premise to manual processes that become redundant pretty quickly. Consider server provisioning. In a data center with hundreds of systems, pushing out operating systems, applications and software updates on a one-administrator-per-one-machine basis gets old fast.
“IT managers can easily identify the things they work on far too much and use automation to ensure that once the job has been accomplished successfully by them, the software can take over and repeat their work” says George Hamilton, director of Yankee Group’s Enabling Technologies Enterprise group.
2. Earmark consistent problem areas
Another winning candidate for automation is the nagging issue help desk operators keep hearing about from users.
Marvin Stone, CIO at New Century Title Insurance Co. in San Diego, says he evolves his use of Opalis Integration Server as his user community demands. He says he regularly checks in with the help desk on the top five complaints from users and instruments Opalis to address issues as they crop up.
For instance, Opalis will alert sales staff that their customers may be experiencing problems when third-party services on a Web site have failed. The software also can post a notification for customers on the Web page that explains services may be affected because of the failure.
“It’s a way of getting ahead of the customer calls and improving service,” Stone says. “We look at Opalis as our Swiss Army Knife. We try to solve a lot of problems by using that automation first.”
3. Establish standard processes
While automation tools promise to cut out manual labor, the technology still requires that you tell it what to do.
Effective automation technology works best when following a standard process that can be adopted across IT silos, industry experts say.
“You have to have processes that can be easily implemented by software. If you aren’t sure how something gets done or who does it, then there isn’t much chance of your software knowing,” says David Williams, a research vice president with Gartner. He suggests that IT managers start out with easily digestible processes that include four to six steps within one IT domain, then add more steps to the workflow as the technology proves itself.
4. Build a corporate data model
IT managers, along with business management, must foster the creation of a corporate data model to ensure automation is effective across an entire organization. The premise of a corporate data model involves all systems using the same format and language to, say, describe an event and normalize data across systems to ensure automation tools acting on the term “event” would work when triggering actions in different products.
Such a model would enable automation technology to work outside of IT silos and report, analyze and optimize systems based on unified definitions and goals, says Bryan Doerr, CTO at IT utility services provider Savvis in Herndon, Va. Doerr uses RealOps’ Automation Management Platform to streamline operations at Savvis and enable the service provider to deliver better products to its customers.
“Having different terms and different data supported by multiple systems is insidious for businesses. In that environment, automation can only evolve in an isolated sense,” Doerr says. “It can be difficult to maintain consistency among the various systems, but true process automation works across many systems and uses the common data model and constrains changes or events inconsistent with it.”
5. Integrate relevant systems
Integration is key to enable automation. If management software and other systems don’t share data or can’t communicate with each other, then automation will fall flat.
And while in the past, the onus of integrating systems fell on IT staff, today’s run-book and data-center automation vendors promise to be the glue that ties disparate systems together. For instance, BladeLogic introduced last fall its Orchestration Manager software, which is designed to provide run-time event, data exchange and workflow integration between BladeLogic software as well as third-party tools. This type of software runs on top of existing systems, sees data and events, and triggers automated actions.
“The whole industry is moving toward better integration to support automation,” says Paul Stallings, senior manager for provisioning services at Blue Cross Blue Shield of Florida in Jacksonville (see related story). He decided to use BladeLogic because the vendor worked with products he already had in-house from HP and others. “BladeLogic is vendor-agnostic, meaning they don’t seem to care where the data comes from. It has created APIs that let its tools communicate with others.”
6. Measure return on automation investment
To keep automation efforts moving forward, industry watchers suggest you measure time, money and efforts of processes before rolling it out, as well as while the technology is in place.
For many IT managers, justifying the purchase of more software can be a challenge, especially when the software promises to do what staff is already doing. But the ROI can be significant, says Evelyn Hubbert, a senior analyst with Forrester Research. One of the main reasons automation technology should be considered is to drive down costs associated with manual labor.
“If you can measure it, you can prove its worth,” Hubbert says. For instance, determine how many trouble tickets are closed within a certain time frame and compare it against the automated process. Or calculate if fewer calls came into the help desk because of an automation feature in the software. “IT needs to take its focus off of mundane tasks. If you can prove the team is working on innovative projects rather than checking off lists, then automation will take off.”
Previous story: Insurer aims for automation | Return to main: Breakthroughs in automation >
