111 I used Uber for the first time a few weeks ago (I like to walk, sue me). It was jaw-droppingly awesome. I had set it up in advance, which took a few minutes, so when I needed it, I simply pressed a few buttons on my HTC (my companion’s iPhone failed in the same task seconds earlier – haha). Uber told me when to expect a driver (5 minutes) and they were right. They knew where I was and where I wanted to go. We were there in minutes. At the end, I just said “thank you” and got out. Minutes later, I got an email with my receipt. Being an infrastructure guy, this made me consider the essence of Uber: an incredibly efficient, global routing network that overlays commodity hardware. In my particular case, that hardware was a giant black Tahoe. It may seem like a slightly odd comparison, but Scality has a lot of similarities. When the RING was conceived, the customers we talked to didn’t say “we want storage.” They wanted a: “massively scalable and reliable mechanism to route users to their data.” These initial customers, typically email service providers, needed technology that could handle far more end users and data, far more efficiently than their current infrastructure. This became an important design perspective for Scality. Support of millions of users and all their billions of small, medium, and large files became a requirement. Performance and low latency became another. Constant uptime was a third. Sort of like Uber. With a clear sense of the problem and the requirements, the core team built something very new. A product that used peering techniques to increase resiliency and uptime. A product that uses parallelism and scale-out techniques to offer strong performance. A product that is fundamentally built for more than a quindecillion pieces of data (look it up – I had to too) across many physical locations. All in application-level software, to better take advantage of the slew of hardware improvements emerging in performance, density, efficiency, cost. It’s fun, and sometimes astonishing, to see the result when someone tackles a problem without preconceived limitations. Uber did just that. And so did we. (@lleung)
