February 16, 2015 | Written by: Billy Cox
Share this post:
When I gave the keynote at Open Compute Platform (OCP) European Summit this past October, I talked about disruptive technologies. As anyone reading this blog knows quite well, our industry moves very quickly, constantly bringing new disruptions to market. That’s a given—and a good thing. In this case, what was disruptive five months ago is even more so today.
When I think about disruptive technology innovations, the first thing that comes to mind is empowering our partners and customers to customize silicon.
In the fall when I gave that talk, I highlighted two examples. For those customers that have very demanding needs, selected forms of customization have already been delivered for more than 35 designs. This turns out to be a very small set of customers with highly focused applications. For a broader range of customers, the use of field-programmable gate array (FPGA) enables higher levels of optimization than is otherwise possible.
This ability to use customized silicon is a disruption twofer: it’s upturned technology and traditional business models. Anytime you talk about specific uses—an FPGA allows for that customization by-design—you can almost hear the thud of financial baggage tagged “low volume” drop at your feet. In this instance, however, it’s doesn’t. When you accelerate workloads with customized silicon, you can make significant improvements to a broader segment of the market.
At IBM’s InterConnect conference Feb. 22-26, I’m going to talk further about disruptions for cloud type architectures.
At the recent Open Stack Summit, we heard it loud and clear from enterprise customers that service level agreements (SLA) remain an important aspect of cloud, but that’s not what keeps them up at night and prevents them from adopting cloud. Availability does. When you’re running a multi-tenant environment, you’re betting your business on it. Without it there’s no performance to measure.
I’m a firm believer though, that once we start pounding on the stability of the OpenStack itself, and once we understand how to do upgrades, rolling upgrades and roll-back, customers will become more comfortable. Once that happens, I predict we’ll see the next major wave of adoption (a.k.a. disruption). We’re still in that learning, advanced crawl stage now.
The telecom segment has led the development of a whole new set of capabilities for using software-defined networking to embed high performance software appliances such as load balancers and firewalls. We are nearing the point where we can place a firewall in front of every application and customize the settings just for that application (and close all others). Integrating these capabilities into OpenStack makes this capability available to a large audience by default.
That’s why OpenStack is so important. Of course, there are many open projects but none have the scale that OpenStack does. It’s that scale that brings innovation from all different corners of the planet and market segments. I can’t imagine trying to address enterprise needs without something like an OpenStack project to be the flashpoint. This is the first time we’ve had a project on this scale where IT operators and users, much less the programmers and the developers, get together and say, “Hey, let’s go solve a big problem like this.”
This is unprecedented. And it’s disruptive in all the right ways.