Red Hat's Hood on Perpetuating SD-WAN Success

SD-WAN's success is linked to a combination of network effects, according to Ian Hood, Red Hat's chief architect for global service providers.

"SD-WAN is one of the use cases that we see quite a bit," said Hood, speaking for Red Hat, "and probably the most popular."

Indeed, open source has become a huge driver for the adoption and evolution of SD-WAN, Hood told Telco Transformation. Meanwhile, SD-WAN can bring yet greater technology to bear when critical masses of population or locations can be realized in the same region. In this sense, therefore, SD-WAN's success rather depends upon its success.

In part one of this Q&A, which was edited for length and clarity, Hood talks about some of Red Hat's SD-WAN work, as well as SD-WAN's flexibility and capacity for democratizing networks. In part two, Hood addresses SD-WAN security issues and SD-WAN's role in video enablement.

Telco Transformation: Is SD-WAN necessarily a white box technology by virtue of it being a virtualized service?

Ian Hood: No. Not necessarily. Initially, most of the available technologies to do SD-WAN were proprietary appliances, and many of the solutions still are today. There are now some that are coming out that are moving to a white box type solution using general purpose hardware, but the initial implementations of SD-WAN just typically used variations on a Cisco edge router.

They were still proprietary, and they basically put a data center blade inside of that router -- x86 capable -- and then some of the solutions out there today still do that. So it's not mandatory to go white box. If you go look at what most of the operators want to get to now, they're calling it a universal CPE [customer premises equipment] approach, pick your operator of choice, and that's where they would like to get -- from the various appliance choices they've used so far out to using an Open Compute Project type of box at the edge of the network.

So it's not mandatory, but it is the evolution and the goal of the operators to move in that direction. And primarily this has been in the enterprise side of the house that this has been happening, this phenomenon. The cable operators and others who deliver broadband to the home are also looking at Open Compute technologies for home broadband as well, but that one's not as far along in the deployment phases.

---

Want to know more about the companies, people and organizations driving developments in the virtualization sector? Check out Virtuapedia, the most comprehensive online resource covering the virtualization industry.

---

TT: There's this focus on the edge of the network. SD-WAN is especially primed for use in sectors where you have a lot of locations -- a lot of branch offices, a lot of retail offices, that kind of thing?

IH: That's correct.

TT: So going from there, to what extent is there an opportunity to see a democratization of home broadband networks to leverage this mass?

IH: Yeah, the opportunity is there. And this is always the question of, "Okay, so I've got a broadband connection to a home, or I've got a broadband connection to a business location." But as I've said before, SD-WAN for an enterprise, if you want some diversity, generally has some other connection -- either wireless or actual physical MPLS -- to give them that option.

So just like we have with every other access technology that has been deployed to date, operators have opportunistically used those access deployments of fiber, or whatever they've done, to get to the edges of the network to bring on broadband customers at the same time. I mean, take a look at what Google has done in Boston. They are delivering broadband to people's apartments in Boston by jumping on somebody else's network. So it's an opportunistic play to go get broadband where there happens to be some high-speed access available. And that is, in my view, limited to where that's economically feasible to deliver those kinds of high-speed access to the populations.

So that's more urban than small towns and middle parts of America. I'm very happy to see that. It's going to happen much more strongly in the larger populations in the Nordic countries where you've got lots of fiber, you've got lots of spectrum, and other parts of the world that have large masses of population in concentrated areas. But when you look at the broad expanse of a US, it's not going to be as easy to deliver that without some further expansion of the optical transport and other access techniques.

TT: What does Red Hat currently have in the pipeline in terms of SD-WAN projects with partners?

IH: We have a number of public deployments of OpenStack in which those operators are deploying SD-WAN solutions with our partners. We have them in Japan, we have them in the US, and we have them in Europe as well.

So there are people like Orange in France, Telefónica is on the mobile side but they also do SD-WAN, NTT in Japan, and Verizon is one of the largest customers we have that is an OpenStack customer that deploys more than just SD-WAN on our technologies. And we continue to add more customers around the globe as they change and expand the collection of services that they offer on their SD-WANs or on their mobile platforms because in essence most of the customers build a platform based on our OpenStack and then they use that same common platform to deliver multiple types of services to their end customers.

So they don't just go build an SD-WAN just for SD-WAN purposes. They go build a platform for virtualization and then they lay their SD-WAN services on top of that versus their mobile services delivery, etc. So it's not just an SD-WAN play for these guys. It's an architectural framework play.

— Joe Stanganelli, Contributing Writer, Telco Transformation