Empowering the CIO Series

Monitoring everything sounds like a good idea, but the complexity of analyzing all that data will grow out of control using conventional methods. How does one handle that kind of scale? How does Rocana address the really big data?

Transcript

Adrian: Hi, I’m Adrian Bowles with RTInsights, and Storm Insights. We conduct ongoing research into the opportunities and challenges presented by new technologies, and the day to day issues facing CIO’s as they attempt to get the most out of their technology investments.

In this series, we ask executives of emerging and innovative technology firms, how they address specific issues that today’s CIO’s struggle with everyday. Today’s guest is the CEO, and co-founder of Rocana.

In our conversations, you’ve talked a lot about monitoring everything. That sounds like a good idea, but the complexity of analyzing all that data is at it grows by orders of magnitude, will grow out of control using conventional methods. With that kind of scale do you need a statistical modeling approach? Can you tell me how you’re addressing the really big data with Rocana?

Omer: Statistical modelling is really just a tip of the iceberg. If you’ve looked at the stop-gap between classic monitoring, which started as “Let’s look at everything from one application”, to where monitoring and [inaudible 00:01:00] today, which is “Let’s keep as little as we need to, put summarize the data in ways that we think is important.” To where Rocana gets to, which is, “Let’s find a way to efficiently store all the data, and then aggregate it in different ways, and represent it to people visually so that they understand where to look to get at the root cause, to get at the underlying data set.”

In today’s environment, you have statistical models, but you don’t have the underlying data anymore, because the systems are efficient enough to collect it, and handle that. In the world with Rocana, you actually have the underlying data, and you have, not just the statistical models, you actually have machine learning, that’s looking at the patterns in the data and saying, “Here’s a time period where I noticed a spike in activity. I noticed a deviation relative to history.” Which we have, because we’re storing all the history and the data.

You also get pre-built aggregates. Whereas with classic systems you had to think about, “How do I want to look at the data, because once I’ve summarized it, the raw data is gone.” With Rocana, we’re pre-building aggregates a thousand different ways. You’re looking at the data from different perspectives, from groups of hosts, to groups of systems, to groups of servers, to physical locations. Then if you want to look at it in a different way, we can just go back and aggregate the existing data set that we have.

The answer is it’s really a collection of big data approaches, to managing the data center.

Adrian: Thanks, Omer, and thank you for watching. For more videos in this series, please visit RTInsights.com