Aligning computer science research with real-world applications
Patrick Crowley earns two NSF grants totaling more than $2.1 million
There are more smartphones in the world than people, yet little to no networking research by computer scientists includes them.
Patrick Crowley, a renowned expert in computer and network systems architecture in the McKelvey School of Engineering at Washington University in St. Louis, plans to develop a new software infrastructure that would allow computer science researchers to develop networking systems and protocols on smartphones and similar devices with a three-year, $1.8 million grant from the National Science Foundation (NSF). Such an infrastructure would allow for better translation between research innovation and real-world use and commercialization.
Crowley, professor of computer science & engineering, said while most networking research projects are designed to improve how users, devices and servers connect and communicate, most academic systems researchers only use the Linux operating system, while the most-used user platforms are the iOS and Android operating systems. A new infrastructure would create software that could operate over several types of devices and make it simpler for researchers to evaluate their work in these common devices.
“Just as it is easier to search for keys under a streetlight, it is easier to design research projects within the comfortable limits of available research platforms,” said Crowley, who directs the Applied Research Lab. “The goal of this project is to dramatically broaden those limits.”
Recent research publications from top networking organizations and conferences featured just one paper on smartphones in the past three years, which shows that the divide between real-world platforms and research platforms is growing, he said.
“It is difficult to generate killer apps for a networked system innovation if the research products cannot be used by people,” Crowley said. “Over time real-world platforms evolve, and research platforms must as well.”
Initially, the team will focus on developing a reliable and productive infrastructure before focusing on high-performance and scalability and hosting it as a service. Once completed, the infrastructure will be available on GitHub with open-source software licensing and training materials available.
Another benefit to this infrastructure is that graduate students will learn the tools and techniques needed to develop and evaluate these platforms that will carry into future positions in research or industry. Later, Crowley plans to develop and teach a graduate-level course at Washington University that focuses on using the infrastructure to develop and evaluate mobile apps and services.
Crowley also received a one-year, $299,000 grant from the NSF to develop a new communication model based on synchronization. The goal of the project is to explore whether the Internet’s long-standing challenges with efficiency, complexity and trust might be better addressed with a new model of communication.
“Communication models are important because they shape what communication patterns are possible and easy to use,” Crowley said. “If these persistent problems with the Internet have their root cause in the underlying communications model, then modifying the existing Internet implementations and protocols will at best mitigate symptoms and leave the root cause untreated.”
The Internet is based on a so-called transmit-receive model, in which messages are sent between a source and a destination. Crowley’s synchronization model replicates state among a group of two or more parties, and more naturally represents the communication patterns seen in modern systems such as collaborative document editing, group chat apps and multiplayer games.
Crowley and his team will study this synchronization model in relation to the transmit-receive and other classical models.