Application Performance Management

Behind the Scenes for the iPhone 5 Launch at a Large Wireless Carrier

By Jay Patel

It’s quite exciting when, as a Solutions Architect for an IT analytics vendor, you get invited to be in a Network Operating Center (NOC) for their launch of the iPhone 5, as I was privileged to be a part of.

New devices and services like the shared data plans are big potential revenue generators, but they can really tax IT systems because they cause such a huge influx of requests that cannot be anticipated. Nor does it make economic sense to maintain all these resources when they are only used four or five times a year.

What was even more exciting personally was that, for the first time, this mobile carrier had implemented the capability to analyze and correlate how the performance of their Point of Sale (PoS) systems would impact business.  Using Netuitive’s behavior learning technology to analyze IT, application, customer experience and business metrics, the customer was able to understand how IT was impacting the bottom line.  Below is just a snapshot of some of the interesting insight that we saw.

Having modeled the workload of the PoS application from the perspective of the JVMs supporting it, we can see that Netuitive automatically discovered a correlation between how hard the application was working (resources consumed and JVM transaction activity), and one of the key business KPIs (SALES_AMT_GROSS).  This is really what you want to see.  If business KPIs are not “normal” (low), but the workload of the application is abnormally high or low, you know you have a serious issue.

On a totally different tangent, Netuitive also discovered the following NEGATIVE correlation when analyzing two business metrics together: the sales of 4G and 3G phones.  People were rushing into the store trying to get their hands on a new iPhone.  If they couldn’t get one, they would consider other 4G phones (and not even looking at 3G phones).  But as the inventory of 4G phones in the stores was exhausted, people were probably thinking “I’m here anyways, may as well get a phone” – and so 3G phone sales increased as 4G phone sales decreased.  That insight was totally unexpected but proven by the analytics.

Anyways, this week has really helped me (and my customer) appreciate the value of advanced APM analytics applied to Point-of-Sale situation.  It’s truly IT analytics with business impact.

I can see how this would benefit just about any retailer or large Telco.  Since we’ve only been working with this customer for a few weeks, my next interesting challenge will be to help our customers develop expertise in using our Trusted Alarms and leveraging them operationally for proactive performance management of their business as usual days.  Very exciting stuff, but at some point, I have to protect my customer’s “competitive advantage” so they’ll keep inviting me back for big launches.

Until next time.

Jay Patel is a Sr. Systems Engineer at Netuitive, Inc. who works with technology and clients to integrate predictive analytics into their systems.

01 October 2012

Posted by: