#RoadCast2017: @SIOSTech Machine Learning For IT Ops & The Self-Driving Data Center
After our successful #RoadCast visit on the West Coast, we met with a number of East Coast companies to get their take on the latest trends in enterprise IT. We spoke with Jerry Melnick, President and CEO of SIOS, who shared his insights on the company’s machine learning analytics product, SIOS iQ.
Melnick explains that SIOS is taking a different approach to analytics than competing products. Based on machine learning, SIOS evaluates raw data coming from disparate sources across the infrastructure to find patterns of behavior for various components (like CPU, memory, network and storage) and how they relate to one another.
So instead of looking at individual, discrete processes and trying to control them through predefined thresholds, SIOS looks at how a VM relates to a storage device, or how a storage device relates to the network. By evaluating patterns of behaviors and these relationships over time, SIOS is able to identify erratic behaviors, pointing out only what’s actually meaningful and actionable for IT ops.
While other analytics and monitoring systems point to individual problems, leaving IT teams to piece the information together, SIOS boils it all down to a single screen that shows exactly what the problem is, where it resides, why it’s happening and what needs to be done to alleviate it.
As Melnick points out, SIOS iQ is essentially brining machine learning to the data center and IT operations. What’s more, it also lays the groundwork for forecasting and the self-driving data center.
Machine Learning In IT
SIOS iQ bring a completely new approach to how IT ops teams get their information. The company is on a mission to eliminate the siloed approach to evaluating IT ops data, instead combining data from multiple sources and leveraging machine learning to get better and more automated over time, allowing IT teams to scale.
Melnick says that the problem with today’s analytics and monitoring tools is that they only look at discrete problems, giving IT limited information that’s often incomplete and challenging to IT teams to identify how different systems relate and affect one another. Through machine learning, SIOS iQ looks at patterns of data more holistically and in a more meaningful way, delivering precise information with actionable insights, instead of alerts to predefined rules that are often arbitrary. SIOS goes even further to identify the root cause of problems that might span across different systems.
Instead of the traditional reactive approach, SIOS is helping IT teams be more proactive, allowing them to project and model different scenarios ahead of time to see what the real impact would be.
Achieving A Self-Driving Data Center
The idea of a self-driving data center is that it’s able to learn and get better over time, Melnick explains; through machine learning, the system is able to understand what the operational environment looks like and respond to it in an intelligent way. This is where the patterns of behavior over time play a critical role.
The promise of the self-driving data center is to be able to react appropriately every time and to be able to predict outcomes from previous behaviors. It will allow IT to look ahead and focus on planning improvements instead of constantly putting out fires and searching for the cause of the problems.