Adaptive Flash Storage
Varun Mehta and Umesh Maheshwari founded Nimble Storage based on an observation at the foundation of data storage – the physical media itself. Six years later Nimble became a multi-billion dollar publicly traded company, and along the way helped transform an industry.
I first met Varun when he was VP of Engineering at FastForward Networks, which was one of our portfolio companies at Accel in the late 90’s. Varun had come to FastForward from NetApp, where he was a critical early engineer. At FastForward, Varun joined a phenomenal team that subsequently went on to great things around the industry, most notably the formation of Riverbed Technologies (but that is a different story!). FastForward was acquired relatively quickly by Inktomi, the dot-com bubble burst shortly thereafter, and Varun subsequently got back into storage and led development at a string of companies including Data Domain. It was there that Varun teamed up with virtuoso technologist Umesh Maheshwari. Umesh is a soft-spoken, incredibly gifted engineer with rare excellence in both high level architecture as well as the details of implementation. Umesh and Varun considered several startup concepts together over the years before finally deciding to take the plunge with Nimble.
The founding epiphany of Nimble was that Flash Memory had the potential to transform enterprise storage, and that taking full advantage of Flash required a totally different technical approach. Varun and Umesh knew from their years of experience at NetApp, Data Domain and elsewhere that the foundation of any storage system is the manner in which data is written to the physical media – the data layout itself, and that this layout was in fact the absolute hardest thing for a storage systems vendor to change. They argued that this fact gave Nimble the opportunity for substantial innovation, at a layer where the large incumbents would be hard-pressed to respond. The result was Nimble’s unique log-structured file system, highly tuned to the nuance of Flash media and the key enabler of Nimble’s hybrid (Flash plus Disk plus DRAM plus even as yet undreamed of media) storage architecture. The implications for customers turned out to be profound.
Nimble’s next stroke of genius also included a stroke of luck. From the very beginning Varun and Umesh insisted that Nimble would utilize a channel-driven go-to-market strategy. This had worked for Data Domain and early iSCSI leader EqualLogic, and would allow the company to gain strength in the mid-market before attempting to compete directly with EMC and NetApp for the major accounts where they were strongest. Nimble’s Sales, Marketing and Product Management leadership (Mike Muñoz, Dan Leary and Ajay Singh) did a superb job of aligning product and distribution strategies around the channel model, and their efforts received a boost from the outside when Data Domain and EqualLogic were acquired in by EMC and Dell, respectively. Key storage channel partners were suddenly up for grabs and in search of new products, and the Nimble team aggressively took advantage of the situation. Today the Nimble channel is a powerful competitive weapon.
While great strides were being made in product and go-to-market, the company’s real competitive advantage – its team – was being carefully assembled. Early hiring was excruciating as Varun and Umesh refused to lower their quality bar. This meant that the company was severely understaffed in its early days, but the pain proved to be worth it as the nucleus of a truly special development team came together.
Team Nimble Getting Started in 2008 (Umesh, Sanjeev Patel, Rick Lund, Varun, and Tomasz Barsczak)
Varun and Umesh had always planned to bring in a CEO when the product was ready to take to market, but once this search was underway it became apparent that it was going to be very difficult to identify the right candidate. Fortunately, outside events once again came to our aid, as Harmonic acquired another of my portfolio companies, a video infrastructure company called Omneon. About a year earlier I had helped convince the CEO of Omneon, Suresh Vasudevan, to join our Board at Nimble. This made sense because of Suresh’s deep storage experience -- I had previously recruited him to Omneon from NetApp, where he had run all of product operations. With Omneon’s acquisition, it was now possible that Suresh might be available to lead Nimble. The founders and team were thrilled with this prospect, having already worked well with Suresh for a year as an outside director. We were eventually able to convince Suresh to take the reins as CEO, and his unmatched leadership and strategic gifts have made all the difference in guiding Nimble from being a “hot box” project to an enduring independent company.
The Nimble story is not without its hair-raising moments, which is part of what makes it such a good one. The initial Nimble product concept was a network storage cache, intended to accelerate access to data on other primary storage systems. It was supposed to allow Nimble to gain market credibility and initial penetration before fielding a more strategic storage array. However, once this first product was in beta, the team realized that its benefits were fairly narrow, and the market for it would be limited.Varun, Umesh and the team took the surprising and courageous position that the cacher should be killed immediately, and all development resources devoted to the storage system you see today.As investors and board members, we were taken aback. The historical success rate associated with such a move was not good.But the team made their case, stuck to their guns and eventually convinced the rest of us. Rarely have I seen a team with such intellectual honesty, forthrightness and conviction!
From that key decision point forward the Nimble team executed beautifully. Revenue consistently exceeded targets in what became a running joke at Board meetings with ritual tongue-in-cheek accusations of sandbagging. Product value was deepened with the addition of multi-dimensional scalability (Scale-out, Scale-deep, Scale-up) and the breakthrough InfoSight support and analytics product, which has set the agenda for the industry. The rest of the world eventually took notice of this company’s quiet but inexorable march, but by then Nimble had built powerful advantages in terms of both technology and go-to-market strength.