Larry Lancaster is proud to call himself a data nerd.
Lancaster, Nimble Storage’s chief data scientist, has been around long enough to see Big Data, the catchall used to describe the massive amounts of data – structured and unstructured – being collected, analyzed and applied to real-life business problems go from theory to reality.
Lancaster, who holds about a half-dozen patents in data use for automated support systems, is the mastermind behind InfoSight™, the innovative engine that monitors all Nimble arrays, collectively and individually, from the cloud. It’s included with every Nimble array.
InfoSight is a singular example of Big Data’s use as a practical business tool. It automatically gathers millions of data points per array each day, making sense of them in real-time with stunning results. Nimble customers are seeing a collective system uptime of greater than 99.999 percent. That means an individual system suffers 5 minutes – or less – of unplanned downtime annually.
Recently, InfoSight became “smarter.” Not only can it foresee potential problems, it can recommend fixes – and it does that in real-time.
How much smarter can InfoSight become? Lancaster, whose LinkedIn profile lists him as “always working on the `Next Big Thing,’ ” is convinced it has limitless potential: “We’ve just scratched the surface.” As an example, Lancaster says InfoSight will soon be able to:
- Fix problems across arrays residing in different physical locations.
- Integrate with third-party software tools and programs to analyze their effect on storage infrastructure.
- Pinpoint and diagnose performance issues in the network and the server.
It’s Nimble’s early commitment to an automated storage management system that makes dramatic improvements, such as these, possible. “To get it right you need to be able to do this at the inception of a product,” Lancaster noted. “There’s no other company that’s built telemetry into its system. If you try and do that later on, you run into performance problems.”
Lancaster joined Nimble in May 2011, not long after it shipped its first array. His mandate: To apply his Big Data expertise to the problem of storage management. Lancaster had previously built an analytics platform for NetApp’s Engineering Informatics team only to find that customers and support derived as much – or more – value from it as engineering did. “That’s where I cut my teeth on different approaches to getting the value out of data,” he said. “It gave me the idea for InfoSight.”
Lancaster soon realized it would be impossible to build a sophisticated storage management system at a legacy storage vendor. Nimble “recognized early that if it collected enough operational data about how products performed in real-life situations, it could analyze that data and dramatically cut support costs,” he explained.
Without InfoSight, Nimble would be forced to investigate reported performance issues by recreating a customer’s IT environment in a lab, an inexact, lengthy and expensive process. “A case that might have taken two or three days to resolve can be taken care of in hours, minutes or automatically.”
But InfoSight has become much more than an automated support system. The insights it’s gathered have informed important improvements in Nimble’s product line. “We understand the workloads our customers are running so that we can anticipate the algorithms and even the parameters for those algorithms to make sure the right data is in flash.”
The amount of data InfoSight has collected is the equivalent of “thousands of system-years of real-world usage” information. That level of intelligence has allowed it to become an invaluable storage-lifecycle planning tool for IT executives. For example, InfoSight can predict when an IT manager will need to purchase a new array long before he has to provision existing gear to acquire more capacity.
Lancaster’s eventual goal: “To understand the lifecycle of a volume – how it grows, how it lives, what its performance characteristics are, and then how it gradually falls into senescence.”
“We have a better understanding of the way storage works in the real world than any other vendor,“ Lancaster said. “That’s because we have the data. It’s all about looking at data as the source of truth.”
- Matthew Miller