Anomaly detection with HPE InfoSight: here’s how it works

[Ronak_Chokshi]

Don’t wait for application anomalies to trigger a disaster for your business. Did you know you can turn to HPE InfoSight for anomaly detection for enterprise application workloads?

If you’re in IT, you’ve no doubt seen anomalies occur in your enterprise. They can come from anywhere and can potentially take all your business-critical applications down in no time. The impact of anomalies can be catastrophic for your business – creating an immediate churn in users leading to a loss of reputation, along with firefighting and finger-pointing within the organization.

Anomalies can occur for a variety of reasons. For instance, an unexpected back-up application elevating read/write activity on storage systems or too many workloads on a single VM or insufficient bandwidth on the network or hardware that’s misconfigured. And the most notorious of all that we find more and more of these days is a cybersecurity attack such as ransomware.

Once an application anomaly turns into a disaster, several recovery solutions are available in the market today, including the industry-leading proven solution from our friends at Zerto, a Hewlett Packard Enterprise company. The Zerto platform incorporates innovative technologies such as continuous data protection that help you recover data and applications all the way up to a few seconds before the disaster.

The big questions is: How do you avoid disasters in the first place?

That's followed quickly by: How do you build a modern digital version of the Walls of Constantinople, behind which enterprise IT teams can sleep freely at night?

Such a digital wall can only be built if it has the right detection mechanisms. I’m talking about one that is built with data collection and analysis as well as with the right AI technologies – and one that constantly observes application patterns. The digital wall should learn from the read/write activities as data moves up and down the technology stack, and alert when the application workloads begin to show anomalous behavior.

I wrote about topic in the first of my under-the-hood blog series where you can read real-life examples of infrastructure issues and our data science-driven approach to solving them.

Focus on the anomaly detection process with HPE InfoSight

The most important piece in building a detection mechanism of this kind is the process itself. The process should include the detection itself. This involves using the right machine-learning and alerting techniques, explaining the anomalous behavior by classifying it correctly, and incorporating feedback from all possible channels for continuous improvement.

Anomaly detection for enterprise application workloads is yet another testament of our continued innovation with HPE InfoSight. It perfectly aligns with our vision of making infrastructure invisible for our customers and helping them pivot their IT operations from being infrastructure-centric to application-centric.

Emily Cheng, our data science lead at HPE InfoSight, and I recently recorded a video to show how all of this works. Check out this video on anomaly detection.

Then please feel free to try out the capability by logging into the HPE InfoSight portal.

To learn more about the industry-leading AIOps platform that powers the HPE GreenLake edge-to-cloud platform, visit HPE InfoSight.

---

Ronak Chokshi
Hewlett Packard Enterprise

twitter.com/HPE_Storage
linkedin.com/showcase/hpestorage/
hpe.com/storage