Discover PerformanceHP Software's community for IT leaders // April 2012
Predictive analytics: See your next storm coming
Analytics aren’t good enough. You need predictive analytics that take your service model into account, to see your next crisis in time to prevent it.
In today’s world of cloud and virtualized IT environments, it’s harder to get complete visibility into the health of your business services. But that means it’s also more important than ever because without that visibility, it’s impossible to achieve the cost savings and agility boost that these technologies promise. IT is under tremendous pressure to help realize those benefits.
Managing a dynamic infrastructure will take more than just reacting to business service problems as quickly as possible. It takes more than manually updating static thresholds, which are problematic to maintain and too difficult to set accurately in the first place. To make the most of your dynamic infrastructure, you need advanced notification of problems so you can solve issues before they affect the business. That’s why IT leaders are talking more and more about the power of predictive analytics. The right predictive analytics tool can help them improve service uptime and performance, thereby y increasing business-generated revenue and decreasing maintenance and support cost.
The best predictive analytics tools enable you to do three things: anticipate, prevent and remediate.
A predictive analytics tool built on top of a real-time, dynamic service model can correlate metric abnormalities with the underlying topology. This correlation, along with advanced analytics and sophisticated algorithms, can let you forecast problems and prioritize issues based on their business impact.
A predictive analytics tool should be able to analyze current and historical data to generate advanced notification of future events, even as much as 30 minutes in advance. By correlating this disparate data—irregular behavior, topology changes and past issues—your tool can leverage a broad base of knowledge to resolve an issue and get smarter at preventing similar issues in the future.
Predictive analytics can make operations management easier by automatically learning the real thresholds in your environment. They should look at historical trends and seasonality patterns over time to establish a baseline of what normal behavior looks like. With real thresholds in place, you can then filter out alert noises and only receive notification of issues that deviate from acceptable baselines.
If you have abnormalities related to a specific service, you should be able to see the related SLAs and know the impact that an anomaly could cause. This way, you can focus your attention on the issues that are most critical to the business. If a problem has occurred before, a predictive tool should also tell you how you solved it previously—saving you precious time.
An ideal tool would help you identify the root cause of a problem so you could take action before it has any adverse effect on the business. Look for an automated event-to-ticket closure remediation process. For example, the tool should notify an event manager and initiate a set of automated workflows to fix the problem.
Think beyond issue identification
In an increasingly virtualized infrastructure, you need better visibility into how your applications and business services are correlated with your dynamic infrastructure, so you can track anomalies across the complete IT stack: network, servers, middleware, applications and business processes. The best way to do that is with a tool that not only identifies problems but actually looks into the future and predicts them, then learns from past mistakes even better than your team could. Without such a tool to illuminate your business services, the cloud turns from a cost-saving prospect into a very dark place.
For more on predictive analytics, check out the white paper on the predictive power of HP’s Service Health Analyzer and the article, “Service Health Analyzer: Decoding the DNA of IT Performance Problems.” Also, APM Digest talks about the fine art of “Forecasting Issues in your Data Center.”
Diving into disruptive technology trends like cloud, mobile, and Big Data, HP’s CEO talks about moving not just IT, but the whole enterprise, into a new era.
Dig into strategic trends with our new Discover Performance Weekly video series, and go backstage at events like RSA.