InfoSight: From Hallway Conversation to Kick-Ass Big Data Analytics
It was a dark and narrow hallway. The conversation flowed in torrents . . . OK, OK! It wasn’t that melodramatic – and it certainly wasn’t fiction.
Larry Lancaster, our chief data scientist, has earned a reputation for being at the center of whatever is “The Next Big” thing. When I joined Nimble Storage in late 2009, I knew we had something big on our hands. And, I knew we’d need Larry.
At that point, we were working furiously to get version 1.0 of CASL™, our Cache-Accelerated Sequential Layout architecture, ready for Beta. I was Nimble’s first support guy; we had little need for it in those early days. Finished arrays wouldn’t ship for almost another year.
But, that was the perfect time to build support into our products. The “secret sauce” of Nimble’s storage management system is something we call the “stats library” – the repository for data we collect each day from all Nimble arrays, individually and collectively, from the cloud. It’s an up-to-date and detailed record of how our arrays perform under particular workloads.
Our engineers came up with the idea for a tracking system built into every array. That system has allowed us to gather millions and millions of data points without impacting array performance.
This became the foundation of our big and deep data approach to storage support and management. Once we shipped the first version, it was clear we’d soon have a veritable gold mine of data. It was time to call Larry.
I worked with Larry before. He’s immensely talented, especially when it comes to the ins and outs of Big Data. When Larry first started at Nimble, his assignment was twofold: first, he needed to get all this configuration, event, and sensor data into a scalable system; and second, he had to build proactive support automation using that data.
Larry got the job done quickly, just as I knew he would. The results were immediate: after introducing our automation capability, we better than halved the number of cases handled manually by our support engineers.
Back to the hallway: in fact, Larry and I were strolling through Nimble’s headquarters when we began discussing the potential of all that deep data now at our fingertips. How could we make use of that to be even more proactive and predictive? How could we use it for more than automated case creation?
The answer: InfoSight™, what would become Nimble’s radical new approach to support and managing the storage lifecycle.
I assigned a few engineers to build a proof of concept. Once finished, we showed it to Chief Executive Suresh Vasudevan. Suresh loved the idea (although he was less than impressed with Larry’s user interface). After a bit of polishing, we presented it to the board. We left that meeting with the go-ahead to double down on development.
In a relatively short time, InfoSight has evolved from an internal tool for our technical support engineers to an essential tool for Nimble customers. Our “deep data” approach to support and proactive wellness delivers insights that allow customers to streamline maintenance and accurately forecast storage needs. Using features such as granular alerting and accurate forecasting, they are achieving a level of overall storage health that vastly improves their productivity and optimizes the availability of their systems.
Recently, Nimble released a correlation analysis view that identifies the root cause of performance problems even when these exist beyond the physical boundaries of an array. It’s only one of the new features we’re adding each month, capabilities that customers can access immediately as they’re delivered via the cloud.
The end result: our customers achieve five nines availability, the gold standard for system uptime.
What’s next? I’ll answer that by telling you that our great InfoSight engineers are mighty busy executing on a kick-ass roadmap.
We’ve been listening closely to customers, and here’s what we’ve heard: they want InfoSight to support analytics and correlation with other systems in their infrastructures. And, they want to integrate InfoSight data into their own reporting and monitoring tools.
Without divulging any details, let me just say then that enhancing wellness health checks remains our top priority. We owe it to customers to keep their infrastructures free of any dark and stormy nights.
Tagged with:
•
•
•
•
•
•
Taken from:
InfoSight: From Hallway Conversation to Kick-Ass Big Data Analytics
