Gordon leads the teams who define and validate functionality for use over the air interface within AT&T's wireless network. He has responsibility for the Radio Access Network, Subscriber Product Engineering and Measurement platforms. He is also previous chair of the Small Cell Forum.
We caught up with him exclusively to find out more about AT&T's progress in virtualization, new opportunities in combined licensed and unlicensed spectrum, and some of the potential applications of 5G.
How has AT&T's technological focus shifted in the last 12 months?
Over the last 12 months we've continued with our network evolution. We started the journey to Virtualize our networks a few years ago and have a target of 55% of our network functions being virtualized by the end of 2017.
In addition to that as we look forward to 5G, we've been doing a tremendous amount of work in the densification area of our radio networks, creating a centralized RAN architecture that allows us to upgrade capabilities and technologies as they become available in the future. So really the last 12 months we've been continuing to evolve the network towards what our target architecture for 5G is, and we're well on our path.
As we focus on 5G, our target architecture is very much a virtualized one - we're driving towards open interfaces, and a centralised Cloud-RAN. We've already made a lot of progress in the core and now we're pushing to virtualize components of the RAN while also introducing edge compute capabilities as well.
What are the greatest challenges facing carriers at the moment in terms of meeting network demands?
Well the way you meet demands on the network is a three-legged stool; we add spectrum, we densify our network and we add spectral efficient features into the network. We've been executing on all three of those. We have deployed incremental carriers across our network. We have been densifying our network with centralized RAN capabilities as I mentioned earlier. And, we've deployed several of the LTE-A features which provide spectral efficiency pretty broadly across our network. Right now it's all about maintaining the balance between device capabilities with the capabilities deployed in the infrastructure to meet the onslaught of traffic driven primarily by video.
Can you tell me a bit about AT&T's plans for using LAA? What are likely to be the deployment scenarios, and what obstacles do you foresee?
We've stated that as we densify our network by building C-RAN capabilities, we intend to add LAA to the pico layer that we are deploying. We've actually gotten some pretty impressive results out of our initial testing in San Francisco, where we achieved speeds of more than 950Mbps. There are multiple markets that we expect to have LAA launched in this year with deployments occurring as we speak!
Most of these deployments will be areas where we've had capacity triggers. This gives us the ability to utilise both licensed and unlicensed assets together, allowing us to provide incremental capacity while delivering on capabilities our customers are looking for.
How do you think new technologies in unlicensed spectrum will impact the telecoms industry?
When you take unlicensed spectrum and couple it with licensed, you actually get a more efficient utilization of all aspects of the spectrum, both licensed and unlicensed. So from a spectral efficiency perspective, we're able to leverage both together, which gives better performance to the end customers, in terms of high speeds and the capacity to serve their sizeable appetite for video.
What do you think are some of the most exciting potential future applications for 5G and next generation networks? (ie V2X trials in San Diego, edge computing?)
When you think about future applications for 5G, the imagination has not even really begun to think about the capabilities that will ultimately be possible with 5G. From an LTE perspective in previous generations we've done a lot to increase speeds. In 5G, you start to take those speeds and include lower latent applications, and that starts to open the possibilities.
If I look back at previous technologies like LTE, and the capabilities that exist using that network today, people hadn't dreamed of most of these 5-10 years ago. I think the same will hold true for 5G - as you build the architecture and you enable capabilities, it has the potential to revolutionise multiple industries by introducing capabilities that they will be able to leverage.
Some things that are already being thought about, from a consumer perspective you can imagine video content being delivered in new maybe Virtual Reality types of ways. Perhaps in the enterprise segment, Augmented Reality types of applications will emerge. There are also opportunities in the Industrial IoT space, healthcare with remote medicine and of course V2X capabilities which are often talked about.
In the end, we will likely see variations of the above applications take hold but the real game changing applications are likely not even being discussed yet which is what makes the evolution so exciting.
What will these applications require of the network?
It really depends on the application, but most of them will require lower latency, and the bandwidth to support the speeds required. So while speeds have been going up every year, in order to sustain speeds with low latency, we have to continue to get more efficiency out of the network. We're also working to get those applications closer to the users, which is where edge computing comes in, to really get those lower latency experiences.
How do you think an OpenSource approach will help support and enable next generation networks?
To understand the value of OpenSource, I like to use a smart phone analogy; the smart phone itself wasn't really the compelling change for users, it was the applications and the app stores which allowed many to contribute. By creating a platform that everyone could contribute to, it evolved and accelerated capabilities much faster; the same is true from a network point of view when you start looking at OpenSource. If you have multiple people that are contributing to OpenSource type of capabilities, that should accelerate the introduction of capabilities needed to run/manage our network. We're focusing a lot of energy with ONAP, the Linux Foundation project that we're very active in. We believe that as you virtualize the network, you need to very quickly create functionalities on top of it to help automate and manage the functions, and that's where you get the full value of a virtualized software defined networks.
How do you envisage the role of Tier 1 carriers in the US changing in the next 12-24 months?
Well, that's a hard one. I really think our roles are continuing to evolve based on the demand of our customers. Really I think for all of us it's about evolving the capabilities of our networks to enable things that change our collective customer's lives. So it's not necessarily just one thing, it's just an evolution of our capabilities that ultimately are going to be the enablers driving the things that aren't even imaginable today. We have the potential to introduce capabilities in our networks which could revolutionize the abilities across several vertical industries.
What would you say to someone considering attending SCWS Americas in December?
The SCWS series is a really great group of events, and I think Avren do a brilliant job of bringing together industry parties to have real conversations, which is what I always look forward to the most.