For our latest discussion, I chatted with Chad Thompson of Verizon about innovation and transformation.
For more articles in this series, please click here.
Prayson: Chad, I'm very glad you could take some time to talk with us about the universal CPE (uCPE) project at Verizon. Could you tell us a little bit about how the project got started and what were the drivers?
Chad: We’ve been looking at this market since 2015 and implementing virtualized services for our customers in the cloud. Through our “voice-of-the-customer” feedback, we recognized that cloud was not enough. They were looking for on-prem deployments in their own data centers, at their own sites, to do complex service changes. Some customers told us that the cloud was too far away and that they don’t want everything hosted there. Most want the ability to run their security policies on site and then create a service chain with an SD-WAN solution before traffic goes to the cloud for additional functions. Customers spoke and Verizon listened. We came up with a model where we could offer virtual services in the cloud and on-prem in a universal CPE deployment.
Prayson: You mention the cloud and that refers to Verizon's Virtual Network Service or VNS. Universal CPE is an extension of VNS, correct?
Chad: It is.
Prayson: How does uCPE complement VNS? You said that customers wanted security on site and some other things that are impossible with a central distribution. That's not to say they didn't want the overall cloud-centric approach. This is a way to give them the full benefits of the cloud, right?
Chad: Yes, it's a complementary service. We don't recommend all cloud or all prem. I think in most cases, customer deployments need both. They need some of that hybrid architecture and solution to achieve their business outcomes. Latency-sensitive applications, such as unified communications, sometimes need processing on-prem and require universal CPE to complement a centralized, broader function. It may not make sense to deploy advanced security functions on site. Those kinds of deployments are better implemented in the cloud, and you should be sending your traffic there. In addition, customers have their own unique niche applications that have to run on site. They're looking for a consolidation story, a simplicity story. Really, they want the peace of mind in knowing that their investment is not stranded like when they went with an onsite, purpose-built piece of hardware. They're locked into that vendor, and we're back to the '90s. You made a five million dollar CAPEX investment, you crossed your fingers and hoped for the best. You were stuck.
With virtual network functions (VNFs) and our universal CPE offer, the message is we don't know what the future holds, we don't know what incumbent is going to screw up, or what innovator is going to disrupt the market. You want that flexibility to be able to adapt very quickly. Instead of living with the pain of your choice, you can now have the flexibility to deploy new services or replace existing services with a tailored approach. At the same time, we are lowering the total cost of ownership with more benefits and an overall better customer experience.
Prayson: Chad, you made a great point about not knowing what the future holds. By deploying an open hardware platform, you have created the ability to innovate and roll out services in the future. You mentioned security services and hybrid cloud. What were the other services that you focused on during the initial deployments of universal CPE?
Chad: SD-WAN was definitely the number one pick. Many of our customers want its benefits. Some want the hybrid WAN capability to stay resilient, while others needed to better leverage their MPLS circuits. We reassured them that their critical applications would go over MPLS while everything else, such as non-critical traffic, went over the public internet. There was already a market for customers looking at an SD-WAN solution, and universal CPE was a natural fit. Today, we have customers that are happy with doing basic routing. We offer routing solutions from Juniper and Cisco on that universal CPE. Then we rolled out WAN optimisation as well. Customers want the most efficient network possible. They don’t want huge video files traversing all around their network. Bandwidth is gold; maximizing their network usage is key. We picked those use cases because in 2015, with our "voice-of-the-customer” approach, we surveyed over a thousand enterprises on all sorts of applications. Afterwards, we studied their feedback and started tackling their main concerns.
Prayson: You mentioned that the customers were asking for white boxes. The Verizon universal CPE project had a strong requirement for white boxes and explicitly excluded hybrid devices. Shawn Hakl has been a public voice for Verizon's customers about the request for white boxes. You mentioned they do not want to get stuck with a dedicated appliance and have a sunk cost. Could you expand on that a little bit more?
Chad: Yes. Let's be real. Most customers don't care if it's a red, white, black or blue box. As long as it gets them the business outcome that they're looking for, then it’s fine. What we found was that through a true x86 experience, where I have nothing but cores, RAM, disc, some Wi-Fi and even LTE, we can get a commercial off-the-shelf (COTS) box at a very low price. That way, I can just increase the value that I'm giving to my customers and reduce the cost on the back end. The reality is, if you go to a black box, it's going to be proprietary to the vendor that you bought it from. Again, you lose that flexibility. We actually introduced a grey box when we initially started in 2016 that incorporated some ASICs and some custom processing. There's a use case for that too. We believe that for some customers with extremely high routing requirements, a grey box is the right solution. However, you do pay a bit more; so, it’s a trade-off game. At this point, you have to figure out what you’re really trying to achieve. That is why we have a portfolio of solutions; we are not advocating that one box is always the right fit.
There is a place for purpose-built, there's a place for a grey box, and there's a place for a white box. You just need to identify your personal business goals and then marry them up to the right solution. We have seen that portfolios including our cloud solution make a positive bang in the market. We have the flexibility, the portfolio and the management suite to allow our customers to come in, hit their SLAs, consolidate their apps and have a great customer experience. All of this reduces their physical footprint tremendously and saves them the cost associated with air conditioning, electricity, and headcount. The way that we simplify the entire story has been a compelling event in the market.
Prayson: At ADVA, we agree with you wholeheartedly about picking the right platform for each application. Another thing about which we agree with you is a need for a multi-vendor solution. When the rise in the universal CPE projects started, one of the hard requirements was for multiple suppliers and no vertically integrated offerings. Now, this did add complexity, but in my opinion, it enabled you to achieve the true vision of NFV. How do you see the long-term benefits of a multi-vendor ecosystem, and how did you balance the risk of implementing that versus the rewards?
Chad: Yes, we know a lot about the market, our customers, and the partner community. We know what the company’s strengths are and what they are not. If I know a company is fantastic at delivering SD-WAN solutions, but their core competency is not security, then it behooves me to buy that vertically integrated solution, knowing I'm settling for their shortcomings. With NFV, now I don't have to make those trade-off choices. I can pick the best-in-breed solutions from the industry, integrate it into a fantastic customer experience. From our perspective, by giving the customers an optimum solution across their entire enterprise application stack, I can create some compelling experiences. It comes at a cost from a Verizon perspective to do that integration. We make the challenges of SDN and NFV invisible to our customers.
We give them the benefits, being the best-in-breed applications, the SLA, the management, a fully integrated stack available on-prem, in the cloud or in some type of hybrid, but without any of the pain of integrating that. Whether it be on an open stack based environment or not. Customers have really resonated with this solution. They don't want to make the trade-off choices of a white label application solution, where they really don't know if the security stack has all the certifications they're looking for. Their chief security officer is very opinionated on what security solutions they use. They've been using closed solutions for years, whether it's Cisco or Checkpoint or Palo Alto. They've got equipment stocks and people dedicated to those solutions. They don't want to move away. With NFV, now we give them that option, so they don't have to. They can still get their solution of choice, but now we give them the additional agility, flexibility and total cost of ownership benefits that they would have never gotten had they went with a purpose-built solution.
Prayson: The customer centric focus that you mention is one of the reasons that I think this offering is going to be very successful. Two more that I think are very important, are the focus on zero touch deployment and the support for wireless connectivity. These are both very important parts of the universal CPE project. Tell us a little bit more about how those benefit your customers.
Chad: Yes, at the end of the day it's about removing complexity for our customers. Today we are going through the pain of ordering a box, sending a technician out, having them run a configuration, applying policy, doing some testing, all within a maintenance window that has to be coordinated, plus a truck rolling cost and everything that's associated with that. Now with zero touch provisioning, I can send a box out to a site, and all the end user has to do is plug in the power cord. The box boots up over the LTE and it phones home. Our system recognizes the box and knows what applications it should have running. Those applications get pushed down after two-factor authentication. Now you have a complete service chain running without having to deploy tech at all, without having to wait for multiple boxes to be hooked up, for the cabling, for the schematics. This box is now self-healing, self-testing, self-provisioning and gives you the values of NFV and SDN, with the full management wrap from Verizon, so you don't need high skilled resources to install these boxes anymore. I can provide very simple schematics, where the power cord goes, where your Ethernet connections hook in, and you're up and running very seamlessly.
The value, from our perspective, is that I do everything in software now and can really recognize the benefits of SDN and NFV. I can streamline that whole process and disrupt the market in a very significant way. It's very powerful when we show our customers this capability. Their jaw drops on how easy it is. You can just see them thinking, “instead of ten sites a day, I can roll out hundreds, if not thousands of sites a day.” Only with Verizon do we see that scalability story in the market. Again, we are seeing many people resonate to that.
Prayson: It's a great story for customers. It changes the deployment time for telco services, from telco time to internet time. Being able to get services turned up simultaneously. As you mentioned, this has not been easy. You pursued a very aggressive timeline on this project. When you started there were still a lot of unknowns. You had to have some flexibility in the willingness to solve problems on the fly. That falls into what we would call an agile development process. Tell us a little bit about how this has gone and what were some of the biggest lessons that you've learned.
Chad: It's been a fantastic learning experience for Verizon. We were one of the first projects to go through an entirely agile process. What that meant, from the Verizon perspective, is that we had our “voice-of-the-customer,” our customer advocates are right there with us every step of the way. Instead of writing requirements, crossing your fingers and hoping the right thing popped out in nine months, now we're doing two week sprints, which makes us hugely nimble in terms of responding to new needs and learning more with our customers along the way. We can implement new capabilities at the drop of a dime, just by listening to what the market's telling us would be something that would really differentiate the experience or the business outcome. It has been a learning experience, but the flexibility and for us to be able to prioritize new market needs and implement those quickly with testing, and the quality of the old process but the flexibility of the new. I don't think there would have been another way we could have done it without SDN and NFV.
There are so many unknowns. The market has not matured, and the standards are not there. We're having to invent a lot of this as we go. An agile development process gave us the focus and the collaboration, not only with our partners but with our customers and our internal stakeholders, such that we could create a powerful solution on the fly, including using a DevOps model. This disrupted the industry and it disrupted Verizon. I think we have grown a lot and in the process, we’ve learned a lot. Agile development is here to stay. We are continuing to use that as we develop VNFs and I believe it's giving us a real competitive edge.
Prayson: That's great info, Chad. Anything else you'd like to add?
Chad: If you haven’t taken a look at Virtual Network Services, it is worth checking it out to see what type of benefits it can bring to your organization. We appreciate and welcome any feedback and discussion of how we can help you achieve your business outcomes.
Prayson: I appreciate your time so much. We'll be talking soon.