OCP's Carter Gets Real on Open Hardware

Open Compute Project CTO Bill Carter opened up at the OpenStack Summit on the economics of vendor lock-in, technological culminations of network disaggregation, and video agility.

On May 8, technology executives in suits and developers in T-shirts and jeans descended on Boston for the four-day OpenStack Summit. The Open Compute Project (OCP) stood out as an exhibitor at the Summit for representing the open-hardware movement at an open-software conference. Telco Transformation sat down with Carter in Boston.

In part one of this Q&A, Carter talked about the evolution of network carrier demands for open hardware and open software. (See Carter: OCP Seeks to Broaden Reach at OpenStack.) Now, in part two, which was lightly edited for length and clarity, Carter touches on the real-world impact of disaggregated, open network solutions and open hardware platforms.

Telco Transformation: Earlier we spoke about the Open Compute Project's philosophy on avoiding vendor lock-in. Does this philosophy necessarily preclude a more consolidated vendor environment? Or, to put it another way, what are the real dangers that you see of vendor lock-in?

Bill Carter: Well, I think vendor lock-in is a concern. It has existed for the last 20 or 30 years in our industry. And, by disaggregating the hardware, we are not preventing vendor lock-in from occurring, but we're making it easier to migrate either the underlying hardware or migrate the software stack. For carriers, they tend to like to buy solutions that include bundled services, and there are a lot of smaller companies -- regional providers, national providers, as well as worldwide providers -- that can offer a very specific set of services. So I think the net effect of this is that there's actually more suppliers available to the carriers now that can deliver the solution that they need.

TT: I gotcha. So supply and demand?

BC: Yeah. Our supply chain is actually broadening with a new set of players -- some of which are small companies -- but they do a very good job configuring and delivering a set of services for a regional telecom provider.

TT: So it sounds like, for the telcos, it's kind of a buyer's market right now, or tell me that I'm wrong?

BC: [after a pause] I think that it can be a buyer's market. My hesitation there is that they have traditionally bought from large companies that offer dedicated appliances that are bundled with a set of services, and they could purchase large portions of their infrastructure from a single supplier -- guaranteed that that would work, but also potentially quite expensive. And now they're able to piece together those appliances through a larger number of suppliers. Hence, you open the door to compatibility issues and test issues, and there needs to be an engineering team to put together that total solution or service chain. But there are some benefits in that, in that you avoid vendor lock-in, you have the potential to dramatically drive down your cost, and -- most importantly -- you have the potential to install a technology quite quickly. You can adopt new technology -- both hardware and software -- very quickly.

TT: How is this enabling digital transformation technologies?

BC: So in the cloud service provider segment, there is a lot of new technology that is being deployed into their data centers. Machine learning and the use of GPUs [graphic processing units] for video coding and content delivery are just two small examples. That hardware technology and software technology is being openly shared through these communities, and it allows the telecommunications industry to pick up on that technology and quickly insert it into their infrastructure.

TT: To touch on video content delivery, what are the challenges and opportunities that you're seeing there?

BC: It's a complex question. In the Open Compute Project , we are seeing a lot of GPU solutions being presented to the community, and at our Summit there were several product announcements that were made that are GPU based. I think that, if those solutions find homes within the telecommunications segment, there's an opportunity for them to create new video services that utilize that technology. And I think that we'll see their business being transformed because they have new capabilities that they're installing into their data center -- both at the data center level as well as pushing it closer to the edge.

TT: So it sounds like a lot of this is about agility, at the end of the day, and reducing complexity?

BC: Absolutely.

TT: Do you have any final thoughts on OpenStack Summit?

BC: It's very exciting to be here in Boston this week. OpenStack as an organization was started a few years before the Open Compute Project was started and it's exciting to see how this community has matured and how deployments are now going to production.

They announced that 66% of OpenStack is now into production. We're starting to see a similar trend, albeit a few years behind, where we worked very diligently to create open hardware platforms. We now have dozens and dozens of skews that are in production at our large contributing companies, but we are now starting to see that scale into smaller Tier 2 cloud providers and regional cloud providers.

I'm very excited to see how OpenStack has progressed. I think that Open Compute is going to follow a similar path, and we're going to see widespread deployment of open hardware in the next two years, and see significant market segment share and market penetration for disaggregated hardware, disaggregated switches, and OCP hardware. So I'm very excited. (See OCP's Baldonado on Faster, Better Service Deployments by Telcos.)

— Joe Stanganelli, Contributing Writer, Telco Transformation