But because of the size and complexity of some enterprise application deployments, it can take some time just to incorporate and deploy analytics for the IT infrastructure and CE metrics.  Meanwhile the executive or line of business owner who is funding the project is saying “What’s in this for me?   Where’s my view of the business?” 

An alternative approach that is gaining momentum is to take a “top down” approach where predictive analytics is applied initially to BAM metrics – giving business executives visibility into the “behavior” of their business with composite application health scores.  So you may wonder “well don’t BAM or Business Process Management (BPM) vendors offer this already?”   And the answer is “no” – at least when it comes to applying advanced predictive analytics to the problem. 

In fact, this approach is so innovative it caught the attention of Gartner Research VP Jim Sinur, who posted an entry about this in his blog.  Definitely worth a look.
Here’s the direct link:
http://blogs.gartner.com/jim_sinur/2012/03/26/visibility-is-the-health-of-the-franchise-success-snippet/

And if you want to follow Jim, here’s the general blog address.
http://blogs.gartner.com/jim_sinur

Link to the publication and see the press release here. BSMdigest Transforms Into APMdigest.com

High profile IT outages are again in the news and have become recognized as the defining example of IT operational risk.  According to one survey, computer outages cost American businesses at least $4 billion last year and as business reliance on information technology and cloud computing grows, the cost of computer downtime will mushroom exponentially.  

Not surprisingly, news is made when these outages hit banks whose online banking services are tied directly to their end users’ bottom line. There are also critical outages in institutional trading platforms that don’t make the news, but are even more costly to capital markets’ enterprises.

Concern over operational risk in today’s global banking and securities trading environments has led to even greater concern about process latencies and performance problems in online banking, trading, and payment systems.  Service outages or delays in transaction processing can result in penalties, compliance issues and ultimately in customer and revenue loss. 

And despite significant investments in monitoring tools for applications and infrastructure performance, 60% of IT operations’ executives acknowledge that they can’t identify and resolve performance problems before end-users or the business is impacted.  That’s because there are innumerable monitoring tools being used in IT Operations and Application Support groups, with no cross-silo visibility or correlation.  These monitoring tools also generate more performance data than IT support knows what to do with. The net result: no “big picture” and no warning about problems until it’s too late.

However, there are some notable exceptions to this scenario.  Some of the world’s largest banks, who were also early adopters of large-scale virtualized infrastructure, were the first to deal with the complexity of virtualization management and the deluge of IT and application data that came with it.  They discovered self-learning, predictive IT analytics as central to performance management solutions that could analyze and correlate data and metrics across different silos such as servers, network and storage.  This cross-silo analysis capability was extended to take into account application infrastructure, customer experience, and business activity data.  The resulting sophisticated correlation of the interdependencies between applications and infrastructure enabled them to detect IT performance and application anomalies before they cascaded into larger failures and outages.

It also resulted in the advent of unified performance dashboards that deliver a composite view of the “health” of key applications.  This new ability to have broad spectrum visibility removed barriers that had previously prevented them from achieving their business processing goals.  It is a service-focused approach that provides a means of measuring the performance of business processes across the monitoring spectrum beginning at the hardware level, up through the application performance and resource utilization, into business items and finally the customer experience.  And unlike rules-based management solutions that rely on human guesswork, these advanced, self-learning analytic technologies continuously adapt to changing business conditions to provide accurate indicators of impending performance issues.

Ultimately, predictive IT analytics lets Service Delivery Managers get ahead of crisis management, giving them time to proactively manage application health and reduce operational risk.

Eight of the world’s 10 largest banks are now running performance solutions featuring these predictive IT analytic layers that sit on top of their traditional application and infrastructure monitoring tools, providing a holistic view of performance across business, application, and infrastructure silos.  They analyze and correlate all the data, turning it into actionable intelligence to increase IT visibility, minimize their operational risk, reduce failed customer interactions, and protect their revenue streams.  

To learn more, check out this white paper on how predictive IT analytics is used to assure service performance and availability of global payments systems.

Executives worry all the time about outages to critical applications.  Especially banking and capital markets executives.  They are most sensitive to the performance of applications that are customer-facing or transaction processing, since problems can mean customer and revenue loss.  So these executives seek the “holy grail”: the executive service health dashboard.

Often these dashboards try to provide a whole slew of dials, graphs and meters to show business or customer experience metrics like average transaction latency, average failed customer interactions per hour, etc.  But is this too complex?  Does it tell you at a glance how the application is doing from a holistic perspective?  Maybe not.

Like high blood pressure – the silent killer – just because an application is available and running doesn’t mean trouble isn’t brewing underneath the surface.  Your application may be running fine now and you Data Center may say “trust me, all your servers are up.”  But if latency begins to trend up, will your support team be able to find, isolate, and fix the problem quickly?

What if you had a way to measure the “health” of each IT component in a service, and then had a rolled up score for the infrastructure from a holistic perspective?  And what if you could incorporate business and customer experience metrics into the service “scorecard” as well?  You would have a true, trusted, “composite” health score for your critical application or service.

But you can’t depend on human analysis alone to “score” the health of your application.  You need technology, and predictive IT analytics software offers an objective way to do this.  The software leverages advanced mathematics and algorithms to analyze and correlate IT data in real-time.  Then it automatically self-learns the behavior of an entire IT environment enabling coveted holistic visibility of applications, business metrics, and infrastructure performance across silos and platforms.    The more “abnormal” an environment behaves, the lower its “health” score.

This type of actionable intelligence provides the extra level of service assurance required for mission critical apps, particularly in banking and capital markets, to minimize operational risk, reduce failed customer interactions, and protect revenue streams.

Eight of the world’s 10 largest banks are now in production with self-learning predictive IT analytics software to predict degradations and avoid outages for their most critical applications.  And while predictive IT analytics has been around for years, allowing enterprises to automate manual and rules based processes for physical environments, it was not until the advent of virtualization where these adaptive, self-learning approaches found their home and are now proving to excel in correlating real-time performance data across large heterogeneous IT environments. 

So what’s the number for your critical application?    I depend on my organizations composite health score from Netuitive on a daily basis.  It self-monitors itself and my number is 95.  And that’s a good thing.

Netuitive was one of 250 nominations for Best of VMworld awards in eight categories.  Netuitive won in the Virtualization Management category for its innovative approach to managing performance of critical applications in large heterogeneous environments.   The award marks the third time in the last five years that Netuitive has been named a Best of VMworld winner or finalist.

“We are proud to continue our streak of recognition at VMworld, the industry’s most visible event for virtualization, cloud and application performance management,” said Nicola Sanna, CEO, Netuitive.  “Predictive analytics software is becoming recognized as a must have for performance of applications in dynamic environments.”

Here is how the National Hurricane Center forecasts storms.

 
Below is a link to a whitepaper shows how much of the same math and science is now being applied to manage virtualization and application performance management IT at some of the world’s largest enterprises.  

From Hurricanes to IT Outages: How Mathematics Helps Forecast and Prevent Disasters

By Graham Gillen, Netuitive

For years, application performance management has been about monitoring the availability of servers by focusing on performance metrics or “information” from CPU, disk and memory.  But things aren’t so simple anymore.  

For one thing, the advent of server virtualization adds the need to examine performance metrics from an ever growing number of storage and the network components, which grows complexity.   On top of this, the behavior of all the components is becoming much more unpredictable, creating uncharted territory for even the smartest engineer, let alone a typical administrator. 

Meanwhile, application performance management emerged as the “right” way to monitor applications.  This added even more “information”, including different types of performance metrics and instrumentation, – from transactions, to user experience, to deep monitoring of software object behavior. 

So now we have even more information to really make the engineers’ job of analyzing the data even more difficult.  We have complexity upon complexity – and it’s become impossible for IT support staff to process all this information and really understand how applications are performing, and the dependencies to the performance of the underlying infrastructure. 

How much information are we talking about?  Industry estimates now show enterprises collecting 300% more performance metrics yet they continue to struggle with antiquated approaches based on setting thresholds and creating rules and relationships to try to model inter-dependent behavior of all the components.   And while there are a growing number of application performance management (APM) tools, they lack the analytics muscle or “intelligence” to help users make sense of all the data.  And this is crucial to be able to correlate application performance to the behavior of the underlying infrastructure and resolve or even prevent service outages and degradations.  In fact, Gartner estimates that by 2015, more than 50% of enterprises will need capabilities beyond those delivered by the current APM tools to manage a complex and dynamic infrastructure and application environment.

But a solution is now emerging in the form of advanced predictive IT analytics software, powered by what Gartner refers to as Behavior Learning technologies.  The new approach automates the management process and provides holistic analysis and visibility enabling cross-domain insight and is serving as the “brains” or “intelligence” for large scale application performance management (APM) solutions leveraging existing APM tools.  

These predictive IT analytics software solutions can analyze and correlate real-time performance and application data from tools like CA Wily Introscope as well as all underlying physical, virtual and cloud infrastructure.  They automate the monitoring, analysis, and correlation of millions of performance data metrics consolidating them down to only meaningful alerts.  They are mathematically accurate and unbiased, enabling you to proactively identify and resolve application performance problems before users are affected. 

The result is an enterprise-class APM solution delivering intelligent, cross-domain insight required for managing voluminous data being generated by a growing number of application performance management tools.   It ensures quality of service and quality of experience for critical business applications to protect revenue opportunities, end-user productivity, and customer satisfaction.

Early adopters include some of the world’s largest banks and telcos who rely on Behavior Learning and IT analytics to predict degradations and avoid outages for their most critical applications i.e. online banking and global payments.   Demand is now taking off with Gartner predicting that 40% of the Global 2000 will have deployed Behavior Learning technology by 2014, up from 10% in 2010. 

Below are are two examples of how enterprises view critical application performance.  Typical APM tools provide a myriad of information on specific application performance.  Enterprises seeking a more holistic approach can utilize open, technology-agnostic IT analytics platforms, such as Netuitive, that offer standard integrations with all the leading APM and monitoring platforms to provide a holistic view of critical application performance in large, heterogeneous IT environments.

The two dashboards below illustrate the difference between “information” and “intelligence”.  On the one hand, you have a lot of data from an APM dashboard, which an expert could certainly analyze to try to tell you how an application is performing.  Of course this might only represent a fraction of the charts that the user could be looking at.

On the other hand, you have Netuitive’s service or application dashboard, allows your experts to take all the relevant metrics that define application performance – from IT, end-user experience, or business metrics – and presents the results in simplified health, workload, and (where applicable) capacity indices.  Once all the potential stakeholders in a company understand the analytics behind the dashboard, anyone can tell – at a glance – how even the most complex application or service is performing.

Information-Centric Monitoring Dashboard - (below)

Netuitive Predictive Analytics Dashboard, providing simple, powerful “intelligence” on how applications are performing - (below)

So does all of the old “information” go away?  Of course not.  Once a problem has been isolated to a specific area or IT silo, you still need to look at the performance metric timelines to figure out how to resolve the problem.  And it doesn’t matter if you analyze that data in Netuitive, or in the interface of the source monitoring system.  The point is – this self-learning, predictive analytics is the only way you are going to be able to manage all the information you know have to digest for virtualization and application performance management.

So ask yourself: with all the money you’ve spend on tools to collect data on applications and infrastructure, is the result just “information” overload, or analytics that provide actionable “intelligence”?

Click here to learn more about Netuitive’s predictive analytics platform for end-to-end management of critical application performance.