Application, heal thyself? Easier said than done, when the business system is critical, custom-built and tear-out-your-hair...
A new offering from Optinuity Inc., an IT process automation company in Bethesda, Md., combines management policy, monitoring, run book automation and enterprise job scheduling technology to help complicated business systems fix themselves by treating maladies as the norm.
Autonomic computing is not new. IBM introduced the term in 2001 to describe a paradigm of computing that is self-managed, self-configuring, self-optimized and, hence, self-healing. The IBM approach is hardware-centric and depends on IT standardizing on certain interfaces.
"What we call autonomic policy management is attempting to achieve the goal of autonomic computing, but trying to do that in a single solution that is not dependent on everyone in the world acting in harmonious and orchestrated fashion," Stouffer said.
"We deliver a very practical product that is a subset of what autonomic computing stands for," he said. "We're trying to focus on the business system itself, so in a telephone company, that might be a billing system. In a financial services company, that might be its foreign exchange systems."
Who, what, when and how
Optinuity is among a group of up-and-comers, including NetIQ Corp., Opalis Software Inc., BMC Software Inc. and iConclude, that provide both generic and highly specific software, often referred to as run book automation (RBA).
RBA technology, a term coined by Gartner Inc. analyst David Williams, is geared toward large companies with complex IT systems. It found early adopters in the health care, financial services and telecommunications sectors.
"Run book automation tools orchestrate the automation of IT operations management processes, unifying the communication of the underlying, supporting, IT management tools in support of the process," Williams said. "By looking at the process that you're managing and providing visibility and reporting on that process, you've got better IT operations efficiency."
Optinuity likes to point to the experience of its largest customer to date, a well-known cable company. This company uses a customer relationship management and billing system that locks up after 15 minutes of inactivity. When the system locks up, the user opens a help desk ticket, which is handed up to a production support engineer, who logs on to the billing system, finds the trouble, resets the account, updates the trouble ticket and notifies the employee he can get back to work.
"The real issue is that the lockdown is supposed to be able to reset itself within five minutes, and for whatever reason it does not quite work right most of the time," Stouffer said. "So when these guys get a request to reset, they have to go in and verify that this lockout is one that should be reset and not locked down for legitimate security purposes."
With run book automation, the same issue exists but rather than assign the trouble tickets to production support engineers, the software interacts with the trouble ticket, identifies the relevant information, goes to the vendor billing system, finds the account, tests for conditions, unlocks the thing, updates the trouble ticket and notifies the user. Since all this happens at computer, not human speed, lockout times were reduced and the well-known cable service provider cut costs in half.
Lest you believe a $1 million tax on a systems glitch is peculiar to the cable industry, Stouffer added: "These things exist in every big company. Complex business systems have these little nuances that are handled as though they are exceptions, and so require a lot of handholding."
Optunity's new Oasis product claims to extend this technology by allowing companies to tackle the glitches proactively, or "upstream," by embedding closed loop autonomic processes at the application infrastructure level rather than in the systems management infrastructure.
Take the lockout problem. "Since we know it happens all the time, we decided why not view this as normal and just let the Optinuity software scan constantly for lost accounts, test to make sure they can be unlocked and then automatically unlock it," Stouffer said.
Oasis provides a closed loop wrapper around the billing systems, and "as a result the whole issue has gone away," Stouffer said. "The accounts still get locked, but they get unlocked almost automatically so users are not impacted to the point where they contact the help desk."
Enter the CIO
Caution: Where there's automation --whether it's lightening the load of the help desk people, or cutting free a few engineers -- there are politics, Stouffer said. Taking advantage of the automation may often require policy changes at the highest level (do not elevate the lockout to a help desk ticket) and always requires organization cooperation.
There are other risks, as well, such as people being afraid to automate.
"They have 20 people handholding an application, and they know it is a terrible way to do it, but they are just afraid innately of automating things," because they cannot imagine every possible scenario that can wrong, conceded Stouffer.
The solution, says Optinuity and others, is to build software that allows companies to iteratively evolve their automated models, as they discover more permutations. Says Stouffer: "We put hooks in and allow them to intervene if one of these unknown things does occur, so they are no worse off than they are today."
Let us know what you think about the story; email: Linda Tucci, Senior News Writer