The Cloud Experience Everywhere

Re: Next-Gen Outage Protection: Harnessing AI and Predictive Analytics

By Jordan Lewy

Worldwide Manager, HPE Pointnext Support Services

A major force behind the astonishing rise of cognitive technologies is their ability not just to solve problems, but to prevent them from arising in the first place. In medicine, as an example, we’ve always had reactive solutions – meaning, you take your symptoms to the doctor after they arise. Or from another perspective, we have generalized preventive approaches, such as recommendations for a healthy lifestyle. In addition, some of these recommendations are both preventative and predictive: such as vaccines. You may get the flu shot every year to prevent sickness, yet the vaccine is updated each year to predict possible common flu strains.

IT organizations have been in the preventative business a long time; they’ve always been focused on preventing costly disruptions and outages. But you may not realize how long they’ve been predictive as well. We at HPE, and particularly with the acquisition of Nimble, have pioneered much of the technology that companies rely on to do that. One of the ways we’ve done it is through sensors and data-capturing devices, both hardware and software, that collect information on the health and performance of our infrastructure – latency, software status, memory, CPU usage – and relay it back to the customer via a portal.

GettyImages-838379016_super_800_0_72_RGB (1).jpg

In addition, we run the data through data analytics algorithms in our data centers to detect known issues; if it matches certain predefined signatures, we will identify the risks and notify the customer. If the problem is severe enough, then we will take a step further: opening a support case, so HPE Pointnext can resolve the issue. For example, in some cases, we are able to detect and remediate the issue remotely, but in some cases, we will dispatch an engineer to fix onsite.  

All of this is true preventive and proactive support; HPE has been doing it for many years, and it’s all very streamlined and automated. What artificial intelligence, combined with big data analytics, enables us to deliver is a new, advanced level of predictive protection against disruptions and unplanned downtime.

A whole new level of protection

HPE InfoSight, is an artificial intelligence (AI) platform that eliminates the pain of managing infrastructure. It features predictive analytics, pattern recognition, and recommendation engines to predict and prevent problems across the infrastructure stack. HPE InfoSight’s sophisticated threat detection engines can analyze trends and patterns over significantly longer periods of time to identify potential errors earlier. We can open the customer support case, process it and close it faster than ever before.

But beyond that, HPE InfoSight and HPE Pointnext together can address potential issues across our entire customer support base – before they even come to light in operations. Let me give you an example, a storage customer support story. Storage is on the front line of our InfoSight innovations; it’s typically a big-ticket, mission-critical part of the infrastructure that companies want to safeguard carefully against any interruptions.3PAR Infosight and HPE Pointnext.JPG

Any infrastructure vendor will, of course, put enormous effort into its quality control efforts, and HPE’s QA is second to none. Still, once a device is out there on active duty it’s subject to unpredictable demands and a huge range of software and hardware configurations (and human error), in addition to issues that may sometimes arise. In this case the customer had noticed a disruption in data availability in an HPE 3PAR StoreServ unit. The customer called HPE Pointnext, and we opened an initial support case, quickly identified the problem and fixed it.

Next, thanks to HPE InfoSight’s artificial intelligence and predictive analytics capabilities, we performed deep diagnostics to uncover the root causes of issues, in this case a memory leak that impacted a critical data service process. In a complex system like a storage array, there are thousands of consumers of memory, and when a system runs out of memory any one of them can be the culprit. It’s a real needle-in-a-haystack problem to find the root cause. Our data scientists were able to scan logs and numerous types of memory allocations in the system to understand and resolve the issue.

However, both initiating a support case and diagnosing the problem are common support approaches. The next two steps in the image demonstrate how HPE’s services have evolved to address new territory for customers. The prediction phase – which is where we identify the telltale signs of a particular problem, code it into a signature, and use AI to run it against the whole install base of our systems in order to identify customers that could potentially face a similar risk.

And at this point we are ready to perform automated remediation. We proactively open a case for each  customer, and assign HPE Pointnext support organizations with the appropriate skills, expertise, and experience to talk to them individually. In essence this support experience is dramatically better because it’s highly contextualized and personalized for each customer. In this particular example, we have helped over 180 customers take the necessary steps to mitigate the issue before it could emerge in a performance degradation or an outage.

AI for a healthier data center

I opened this post with a reference to health-care, and another medical analogy comes to mind here: The whole AI-powered process I’ve just described looks a lot like the sophisticated disease control of modern health systems. Your doctor handles the immediate support case and takes care of your symptoms, doing the diagnostics and calling in specialists as needed. Epidemiologists can aggregate data and determine trends and risks, and health policy makers can intervene when necessary to address a threat preemptively and prevent a problem from spreading.

However you want to picture it, the combination of HPE InfoSight and HPE Pointnext is a potent prescription for the health of your data center. And it might be just what you need to bring your IT staffers’ stress levels way down.

Get started today. Read more about HPE InfoSight – artificial intelligence for the data center, or visit our HPE Operational Support Services website to contact a sales team. The HPE Operational Support Services experience will help you manage the day-to-day IT operational tasks while also freeing up resources to help your business stay ahead of the competition.

About the Author

Jordan Lewy.jpgJordan Lewy is a Worldwide Manager for HPE Pointnext Support Services. In this role, his goal is to transform HPE’s customer support experience using HPE InfoSight, which in turn drives their business outcomes and enables their digital transformation journey.

Jordan brings to his position a well-established background in information technology and professional services, where he has worked for over 20 years. Prior to taking on his current role he held other positions in HPE including leadership for HPE’s Storage Support services, Installation and Technical services and HPE’s Customer Technical Training business.

About the Author


HPE Pointnext Services experts share their insights on the topics and technologies that matter most for your business.


Thanks Jordan. The article is excellent.. connecting the dots between HPE Infosight and HPE PointNext ...