It's never good when a server goes down. At an airline like Israel's El Al, it could mean problems for the workers in the operations control center who are charged with keeping planes in the sky. Or it could mean that George in HR is going to have to use a different printer.
Until recently, it was hard to tell the difference. Any IT infrastructure issue triggered an alert, whether it affected the ticketing system or accounts payable. The IT group could not assess the business impact of the problems and some repairs took too long.
The company's Tel Aviv data center had grown increasingly complex--hundreds of servers, increased virtualization, many applications. An IT service management (ITSM) suite enabled the IT team to figure out which technology piece had a problem, but not what the real business impact would be, says Itzik Cohen, El Al's director of infrastructure and communications.
Traditional ITSM tools discover all the components in an IT infrastructure, which are then put into a database, explains Linda Musthaler, principal analyst at Essential Solutions. Then someone has to manually correlate the IT assets to business services. El Al attempted to do that, but mapping a single service took weeks, and by the time the process was done, the map was out of date, Cohen says.
El Al's infrastructure group wanted an automated way to connect the dots between the airline's business processes and its underlying systems. "We wanted to take more proactive action for the business rather than simply monitor systems," Cohen says.
Last year, El Al signed on as one of the first customers of Neebula ServiceWatch, an IT-management tool focused on business processes.
Once the infrastructure team provides an entry point, like a URL or a simple object access protocol call, for a specific business activity--say, the frequent flyer program or plane maintenance scheduling--ServiceWatch detects all associated components. The tool can also map business services to components of proprietary, homegrown systems that were previously considered inscrutable.
The tool models how lower-level events might affect high-priority services: those that generate revenue, like the contact center or website, or those that play a safety or security role. A dashboard displays the health status of critical systems on a single, color-coded screen. El Al has mapped its 60 most critical business processes, with the remaining 170 to follow this year.
Modeling most business processes takes just a few hours. Those that involve homegrown systems take a few days. The hard part, Cohen says, was defining El Al's 230 business processes. That required consensus between not just IT and its business stakeholders but also between IT and IT, he says.
There had always been some friction between the applications group and the infrastructure team about the source of problems--was it a server issue or did something within an application change? Applications, infrastructure and the business had to get on the same page.
"But at the end of it all," Cohen says, "we finally have some common knowledge. It's not just that everything is down or everything is up."
Cohen expects that service availability will continue to improve now that IT can immediately pinpoint what process a specific component will affect. "This enables us to prioritize where and when to spend our resources."