Smart Enterprise Magazine

Volume 8, Number 1, 2014

Issue link:

Contents of this Issue


Page 21 of 23

of the business. Green lights across the board may offer the illusion that all is well, but trouble could be lurking below the surface. With millions of metrics to look at, how does a person or team find the proverbial needle that could be located in one of many haystacks? IT organizations are feeling the pres- sure. In the September 2013 study "Market Pulse: Application Performance Manage- ment," CIO magazine surveyed over 100 senior IT leaders about trends in APM and found that 99 percent of respondents agreed that it's important to have the ability to quickly and accurately identify problem root cause. Yet only 36 percent of respon- dents rated their ability to do just that as excellent or good. Similarly, when it comes to proactively identifying and diagnosing problem root cause, 93 percent of respon- dents claimed that it's important, yet only 32 percent felt confident in their organiza- tion's ability to do so, indicating that they're more than likely reacting to issues after they become customer-impacting problems. This gap between the importance of being able to identify problem root cause and actually doing so resulted in lost employee productivity due to downtime or slow performance (55 percent), spending too much time resolving issues rather than driving innovation (46 percent) and hav- ing the business lose confidence in the IT department (42 percent). It's an uphill battle. Julie Craig, Research Di re c t o r a t En t e r p r i s e Ma n a g e m e n t Associates (EMA) compared the issue of keeping pace with the flow of IT met- ric data to the iconic "Candy Factory" episode of I Love Lucy in the July 2013 webcast "How Advanced Ana- l y t i c s Ad d Va l u e t o A P M " w i t h C A Te c h n o l o g i e s , "If you can visualize the assembly line today with candy stacked s e v e r a l f e e t h i g h , t h a t w o u l d m o r e closely resemble the Big Data operational environments we're s e e i n g t o d a y. T h e product has changed, but the idea is the same: H u m a n s h a v e l i m i t s , whether they're wrapping candy or they're analyzing metrics. Today's massive application eco- systems generate far too many metrics for us to intelligently troubleshoot without real-time analytics." APM: Analyzed To assist customers in addressing the issue of finding problem root cause quickly and to enable them to get more value out of their APM Big Data, CA Technologies added Application Behavior Analytics to its CA APM offering. Application Behavior Analytics uses multivariate analysis to process APM data (up to 100,000 simul- taneous metrics) and look for anomalies. It does this by automatically learning relationships between components and what normal performance looks like over time, so it can quickly determine if certain metrics are out of line. Using Application Behavior Analytics, a government agency discovered its wide-area network router was dropping packets, causing slowdowns in a critical application. Similarly, a large consumer products firm discovered previ- ously unknown performance anomalies in its service-oriented architecture environ- ment by applying analytics to its APM data. The beauty behind the technology is that it requires no input from the opera- tor to know what to look for, so it provides an essentially impartial view of what's happening performance-wise. It also pro- vides a view into application component relationships that IT operators may not have thought to look at in the first place, particularly in large, complex systems with multiple moving parts. "To be able to take the metrics that we're used to going hunt- ing and pecking for, and [CA APM] pops up and says, 'This is different than what normal is and all these things changed a t o n c e ,' " s a y s L i n Richardson, Senior Middle-Tier Admin a t I n t e r m o u n - t a i n H e a l t h c a r e in Salt Lake City Utah, explaining a potential scenario w h e r e C A A P M analytics can show how a change in one a p p l i c a t i o n c o m p o - nent behavior could be linked to similar changes in a database server's memory and CPU utilization. "It's just another tool to help us be more efficient and address increased expectations of the business on IT." It's important to note that the anomalies found by Application Behavior Analyt - ics' algorithms are not always bad. If a customer has added new application serv- ers to improve performance, Application Behavior Analytics will show the change in behavior, verifying that performance has indeed improved. "Applications are the lifeblood for our customers, who are continually being pressured by the business to deliver a great end-user experience for internally and externally facing applications," says Rick Fitz, Senior Vice President of Prod - uct Management at CA Technologies. "But as applications grow more complex, our customers need more sophisticated tools to help them stay ahead of potential problems and solve those that do pop up more quickly. By adding Application Behavior Analytics to CA APM, we're giv - ing our customers an automated set of machine-driven eyes that can help find complex problems early and provide better root-cause analysis to solve problems fast." In today's business world, customer experience, satisfaction and loyalty can be heavily dependent on IT and the applica - tions and business services it supports. With a projected shortage of data scientists that can help process Big Data, according to an October 2, 2013 CIO magazine report "Who's Training the Next Generation of Data Scientists?", IT organizations are going to need as many automated tools as possible to keep ahead of the mount - ing piles of IT operations metrics from systems like APM. "We're beyond where we can throw people at the problem. I really think there's no other answer at this time than analytics and automation," says EMA's Craig."Analytics are the secret sauce that transfor ms machine data across the application ecosystem into actionable, useful information. They mitigate complexity and take lots of info and make it meaningful to humans." ■ JASON MESERVE is Principal Product Marketing Manager at CA Technologies. "We're beyond where we can throw people at the problem. I really think there's no other answer at this time than analytics and automation." 22 SMARTENTERPRISEMAG.COM

Articles in this issue

Links on this page

Archives of this issue

view archives of Smart Enterprise Magazine - Volume 8, Number 1, 2014