When Amir Alon, Yori Lavi, and Israel Mazin decided to form a startup in 2002, they went to some Fortune 1000 corporations and asked what kept their IT pros awake at night. The answer: measuring and maintaining performance as their IT infrastructure shifted from client-server to .Net and J2EE architectures.
“We were astonished,” Alon says. “There were all these big players in the performance management market, but their solutions couldn’t adapt to the new paradigm, where a single business transaction can span many different systems.”
The former Memco and Compuware executives called their New York-based company OpTier, with the goal of optimizing performance across all tiers of the enterprise. Two years later they introduced CoreFirst, a performance management app that enables corporations to analyze activity at every level, from mainframes and middleware to Web and application servers.
With CoreFirst, every IT “transaction” — whether it’s pulling data from a server, requesting a Web page, or sending a job to a network printer — is assigned a unique ID. Software agents installed at every branch of the network follow the transaction as it flows from one system to the next, providing real-time analysis that allows IT managers to quickly identify system bottlenecks. CTO Alon says managers can get as granular as they want, looking at particular types of transactions or drilling down to view a single user’s activity.
Alon says the idea sprang from his days as an IBM mainframe performance analyst in the early 1980s. With Big Blue iron, transaction tracking and workload management were tightly integrated into the OS, he says. “But in distributed environments these concepts didn’t exist, or they existed only on silo systems.”
Now the transactions users are running and their impact on the system as a whole can be made completely visible to the enterprise, Alon says. “You can isolate what part of your infrastructure is causing performance problems.”
More important, CoreFirst allows IT departments to assign priorities to different types of transactions. For example, a financial institution might give priority to its online banking customers, allotting more system resources to them and fewer to lower-priority jobs. This allows enterprises to improve customer service, use their resources more efficiently, and avoid expensive overprovisioning.
The biggest challenge? Building a system without introducing additional overhead or latency, Alon says. But don’t expect the former engineer for the Israel Defense Forces Computer Center to hand over those secrets. “That’s our secret sauce,” he says. “We’ve got patents on that.”