September 10, 2021

Q&A with Edward Oughton, professor at George Mason University and Facebook academic collaborator

By: Meta Research

In this monthly interview series, we turn the spotlight on members of the academic community and the important research they do — as thought partners, collaborators, and independent contributors.

For September, we nominated Edward Oughton, a professor at George Mason University. Oughton has been an active member of the Facebook research community since 2018 and has contributed unique cross-disciplinary insight to connectivity challenges. In this Q&A, Oughton shares more about his research background, the start of his collaboration with Facebook, his research interests, and his current projects.

Q: Tell us about your background in academia.

Edward Oughton: I have an MPhil and a PhD from the University of Cambridge. My PhD research was originally in urban and regional economics, focusing on the economic impact of broadband connectivity infrastructure. I then undertook a postdoc at the Cambridge Judge Business School, again focusing on modeling infrastructure but developing risk analytics for the insurance industry. I then moved to a Senior Research Associate role at the Environmental Change Institute at the University of Oxford. I have also held visiting positions in engineering and public policy at Carnegie Mellon, and in the Network and Systems Group at the Cambridge Computer Laboratory. I took up my current role at George Mason in 2020.

This background is evidently very mixed, given that I spent time in economics, engineering, and computer science. I certainly used to see this as a weakness, but over time I found that these three areas very often intersect, yet there are few researchers adept at working across these subjects in an integrated way.

Q: Tell us about your role at George Mason University and the type of research you specialize in.

EO: I am currently an assistant professor in data analytics and enjoy the fact that I’m given freedom to work very broadly. Other research opportunities often require you to publish in a very narrow set of subject-specific journals, so I am thankful I am not so restricted.

One of my main areas of research focuses on broadband connectivity, including developing and testing strategies that help connect more people to a faster internet. Connectivity challenges are very complex, and to implement sustainable connectivity solutions we need to solve engineering, economic, and social challenges all at the same time.

The industry practice today is built around very simple spreadsheet models, which use very basic assumptions. However, these simplistic models do not address key challenges directly. For example, network design models do not include spectrum strategy considerations, whereas spectrum policy studies often do not directly include engineering choices.

In contrast, my research attempts to develop holistic models that can be used to compare the design of different network architectures as if they were about to be built by a network operator. I then use these estimates to quantify the required investment to connect unconnected users. Such an approach reduces uncertainty compared with spreadsheet approaches, as both network quality of service and the cost of the network can be estimated using a single model.

Q: What have you been working on lately?

EO: Along with Julius Kusuma and Erik Boch, we have been wrapping up a project that focuses on non-line-of-sight diffractive wireless backhaul. The Facebook team engineered the technology and deployed various test sites, with my work then quantifying the wider cost-efficiency savings of applying such an approach across a full network, using Peru and Indonesia as examples. The software codebase is open source and freely available for other engineers to explore and develop.

We have also published a qualitative review where we critically compared 5G and Wi-Fi 6 technologies, which makes a pleasant change from focusing purely on quantitative research. The paper gained quite a bit of publicity and is currently the most downloaded paper in the journal Telecommunications Policy. We make the point that cellular technologies such as 5G are best suited for serving users who require mobility when moving between their home or workplace. In contrast, Wi-Fi technologies such as the new Wi-Fi 6 standard are much cheaper and easier to serve indoor users. One group of commentators claims 5G will eliminate the need for Wi-Fi, but we don’t agree.

More recently, we’ve been pivoting the broadband models developed to help us understand more about the sustainability implications of universal broadband. This is rarely considered but is increasingly important to be aware of. For example, we’re interested in quantifying the environmental and sustainability impact of different network architectures. To reduce this impact, you often need to reduce energy consumption, which can potentially minimize cost, thus helping with the economic feasibility of providing broadband connectivity in many challenging areas.

Q: When and how did your collaboration with Facebook start?

EO: Julius Kusuma read my early 5G assessment work from around 2017, and he invited me to a Facebook Research workshop in Menlo Park. We then discussed possible projects, and I was encouraged to apply for the open call put out by Facebook Research. I submitted an open science research proposal which was eventually funded, and we have gone on to successfully publish numerous papers.

Q: What are some of the research challenges you’re addressing with Facebook? How do you approach these questions and contribute your expertise?

EO: We are aiming to help provide broadband connectivity to everyone around the world. Currently, the costs of building broadband infrastructure can be higher than users are willing or able to pay, leading to a phenomenon widely referred to as the digital divide. Increasing per capita income, and therefore the amount consumers can spend on broadband, is highly challenging. Thus, our efforts are focused on trying to reduce the costs of providing broadband connectivity, which requires new technologies, as well as innovative business models and much more encouraging telecommunication policy and regulation.

Often, there is a lot of qualitative discussion around the problems being faced and how to solve them, in order to provide better broadband connectivity globally. My contribution is to reduce the complexity involved with this problem, to enable quantitative modeling of broadband connectivity options, thus providing comparative analytics that indicate performance. This evidence can then be disseminated to either industry or government decision makers, to help them make better decisions.

Q: Where can people learn more about your research?

EO: I am strongly committed to providing the data, code, and written papers developed through this research openly, so that others can evaluate and improve on the activities undertaken. Therefore, the code is available on my GitHub homepage, and our research outputs can be found on my Google Scholar page. My arXiv page also contains recent working papers.