Prayson: Tim, we have worked closely together for a good while now and I appreciate you taking some time to talk with me today about NFV. Please tell us about Masergy's role in the NFV transformation.

Tim: We were early adopters of NFV, and I think we (Masergy and ADVA Optical Networking) share a vision as to what we want to deliver. Most of our customers are global, as you know. Most of them say, “If you can reduce the footprint in remote locations that’s solving a really big problem.” It’s solving a really big problem for me, too, because I don’t have to ship those boxes and I don’t have to wait for them to get through customs. I don’t have to have them delivered to the wrong place by a courier.

That was really our vision for the Virtual f(n) product. We wanted to create an all-in-one device. We already deploy ADVA and other NIDs in our network, and those are always part of our solution because that’s how we take a traditional TDM service and convert it into Ethernet. Since we already have that device there our goal was, “Can we find a way to put firewall and a router in that as well?”

Most of our customers, especially the ones we get through large system integrators, have all three of those devices sitting locally. Our goal with this NFV solution was to collapse them into one device. We initially called it premise NFV, and then I think you coined the term pure-play NFV, which I think is really great. I started saying pure-play NFV and I credit you when I say it.

Prayson: I appreciate that. To achieve that goal of replacing a stack of appliances with a single server and software meant that you had to solve a number of problems. What were those problems and how did you solve them?

Tim: The first was to find a supplier that shared the same vision. There are as many flavors of NFV as there are definitions of the term “object-oriented.” It has come to mean almost anything at this point, but what we were really looking for was a premise-based solution. We already had a cloud-based solution for routing and for firewall, but sometimes the customer has a certain requirement — local survivability, marking QOS, whatever. They need a router locally.

We found some partners that said, “We don't believe this can be done. We will go ahead and give you a device that has the carrier Ethernet functionality and hardware, and then we'll just kind of tack a PC onto the side of it.”

I didn’t like that approach because I’d still have one foot in each world. Other vendors said this should be done with white boxes, but they were not going to supply boxes. We weren’t quite ready for that. I don't think the industry was quite ready for that, at least in the prem and potentially in the colo as well.

We wanted a vendor that understood the carrier environment and understood the data center environments that we work in and that our customers have an expectation of devices working in. It needed to be low power, and it needed to meet high requirements from a temperature standpoint. Those were some of the initial challenges.

That's when we came across you guys. Then we went through the whole process of, “Can we fit multiple VNFs on a device?” You guys went through some amazing engineering in terms of optimizing the code. We saw an order of magnitude type performance gain when you went to the deep packet stuff.

Prayson: You're talking about the performance of the vSwitch?

Tim: Yeah, the vSwitch itself. That was really the key step. Then, as we started working with VNF suppliers the hard part was getting the business model right. I needed to have a device that was inexpensive enough that I could put it out there even if the customer didn't need VNF functionality today, but may want it in the future. I always use the analogy of if you're going to put a NID in Mongolia you really don't want to go back to Mongolia and put in a router later. One trip to Mongolia is probably enough.

That was one dynamic. Another was the cost. I had these different elements working against each other, and at the same time we were in a very early market stage. When you asked most of the VNF suppliers for the lowest price they would say, "Well, what do you think it is?"

Prayson: Free!

Tim (laughing): No one ever accepted that answer, but we did try. It was kind of an early shakeout while trying to write the business model. We came across some partners (Brocade, Fortinet and ADVA) who wanted to work with us. You all saw the advantage of being first movers in this environment and gaining the knowledge and the learning that we gained through this project. We solved all of those problems and were able to launch on budget and on schedule.

Prayson: You mentioned a couple times the immaturity of the market and some of the different views and flavors of NFV. When do you think NFV will go mainstream, especially for the Tier 1 operators?

Tim: Every presentation I’ve watched or seminar I’ve attended, they’re really focused on the colo solution, or they’re focused on a set-top box solution. The reason we chose the customer prem wasn't because no one was looking there. It was honestly because that’s what our customers where asking for to solve a problem that they had.

I think in the colo you’re going to see some advantages. I don't know from a cost standpoint how those models are going to work out if you’re buying racks of very expensive quad-Xeon boxes running disk arrays. I buy those servers and know how much they cost. They're not a thousand dollars at Best Buy. They're expensive boxes, and if you layer the licensing cost for a multi-tenancy environment on top it’s even more. I haven't been able to work that model out internally here. But that's where a lot of people are focused. Their approach really is more about cost savings and less about solving a customer problem.

Prayson: As you move down the path toward virtualization, what’s the one thing that’s been most surprising?

Tim: I would say how well the VNFs work together (laughs). Honestly, we’ve been told by many people that this can’t be done.

Prayson: It was a surprise?

Tim: It was a very pleasant surprise. What we found is that with the vSwitch there these devices really look like they’re just upper devices on a network. The ability to control that with software within the device works a lot better than any of the naysayers said it would.

I've gone to conferences on NFV, and there’s a big contingent of people that say, “This can’t work unless you buy everything from the same vendor.” I just fundamentally don’t believe that.

I don't believe that closed approaches are the right approach, especially in today's marketplace. We need to be able to move at software speeds, not hardware speeds.

Prayson: You've now got a commercial proof point that you're right.

Tim: Now we're expanding our VNF store. We have an offering coming out later this month that's a light firewall that doesn't include all the UTM solutions. Encryption and WAN acceleration are both in my lab right now. We’ll continue down that path, even looking at some testing and fault tolerance tools that we use in our network today that sit in the colo that we may be able to push out to the customer prem.

Prayson: All of this innovation will be able to be done with the hardware you’re deploying, with your operational systems and with your current teams.

Tim: Absolutely. One of the reasons that we like our VNF partners is that we already use Brocade in our network. We already use Fortinet in our network. When we rolled this out I didn't have to go to the NOC and teach them how to use a different firewall from another company. I was able to say, “This is the same thing. It’s just a different IP address.”

How you get it set up is different, but that’s a one-time thing. How you maintain it going forward and how you interact with it on a daily basis, that's the same. I wanted to deliver it to those brand names for one, but also I wanted to leverage the operational knowledge that I already have in my group.

Prayson: Tim, you have noted that in your career at Masergy you were the CIO before you became the CTO. One of your first initiatives at Masergy was an overhaul, upgrade and general modernization of all the IT systems. How important is that type of effort to other operators as they move forward in virtualization, breaking down silos, becoming more agile, and modernizing?

Tim: I think it’s absolutely critical. I think the OSS/BSS can be an albatross around the neck of the carriers if they’re not careful. I think you have to go back and evaluate what’s working and what isn’t working. The next generation technology we’re very committed to is open source. I’ve been challenged on whether you should do open source or buy a commercial solution. I've done both of those in my career. I've had more people supporting a commercial solution and still had to wait on vendor releases for six- nine months or even a year longer than what I have been able to write internally.

Open source gives you such an advantage in terms of the basics. You can take the basic LEGOs and snap them together quickly and then focus your code writing and your developers on the business logic that’s unique to your business. I wouldn’t trade that. Somebody can come and write me a check today and say, “Go buy an entire new OSS and BSS and buy it from company X.” I would say, “There's no way. I’d rather have my six developers doing it.”

Prayson: Tim, I've been very pleased and proud to be working with Masergy over the years. One of the things that is most valuable to us at ADVA is that Masergy is not only a customer but also a partner. We’ve had cases where you needed innovative features and were willing to take early delivery to verify that they were meeting your requirements, which is what today we would call agile development.

That synergy helps drive us down the road toward innovative features that other operators will want. That’s been very valuable. In this Virtual f(n) release, the thing that’s been very valuable is you're driving an ecosystem composed of three suppliers to work together, both technically as well as commercially to meet the service requirements, the CapEx and OpEx requirements, and the performance requirements for a real service. Then you tie that back into your control systems to achieve the end goal, which is to make services configurable and dynamically available to your end customers.

Tim: Thank you, Prayson. We're very happy with the relationship as well. Honestly, you guys really demonstrated your engineering chops on this and I’m very proud of this product. As a team, these four companies have shown a path to the next generation delivery of carrier services.