NSF supports research on interdependent networks, software protocol stacks-Silicon Valley Campus - Carnegie Mellon University

NSF supports research on interdependent networks, software protocol stacks-Silicon Valley Campus - Carnegie Mellon University

Tuesday, September 16, 2014

NSF supports research on interdependent networks, software protocol stacks

This summer the National Science Foundation (NSF) announced support for two research projects at Carnegie Mellon University's Silicon Valley campus, both addressing key technology issues in modern networking and security.

Osman Yağan, assistant research professor in ECE, was funded by NSF's Division of Computer and Communication Foundations for his work on Robust and Optimal Design of Interdependent Networks[NSF award link]

The three-year, $451,000 award, which began August 1, will provide support for Yağan's research in modeling, controlling, and optimizing the robustness of interdependent networks. These types of systems, such as the smart grid, are critical to 21st century technologies but also fragile and vulnerable to cascading failures between networks. Past research in network science has focused more on standalone networks; instead Yağan will develop tools needed to address vulnerabilities and mitigate risk in modern interdependent networks.

The US is a major consumer of smart-grid technology, Yağan says, and developments in renewable energy, electric vehicles and interconnected devices "point to a future where the reliability of the smart grid will become paramount."

From the abstract: "Successful completion of the project will require the development of new techniques and approaches in the fields of network science, discrete optimization, and random graph theory, together with acquisition and analysis of real-world data from existing smart-grid networks."

The first phase, now underway, studies network performance and robustness under various topologies and conditions, with the aim of revealing intricate trade-offs involved in the design of interdependent systems; research will then proceed to seek optimal resource allocation strategies for maximum robustness. Yağan is planning to work with two PhD students and several master's students throughout the project.

The research team will also develop a simulation tool to test large-scale interdependent networks against attacks and in various failure modes, and research materials will be considered for inclusion in a new course.

In July, Silicon Valley campus researcher Temesghen Kahsai also received three years of NSF funding for a collaborative project on Verification of Heterogeneous Software Protocol Stacks. [NSF award link] The project is in collaboration with Prof. Zvonimir Rakamaric from the University of Utah.

Software protocol stacks — such as those supporting Bluetooth or USB operations — are sets of software that typically connect layers within computers or devices, from hardware to operating system to high-level libraries. These highly heterogeneous and complex stacks are usually developed over several years with many people, products and processes involved, and with so many factors, several errors persist through development and into production, where the cost of fixing them grows.

Correcting them is a "tedious manual process," Kahsai says, adding that these bugs negatively affect the user experience and can present security risks. A more reliable protocol stack, he explains, requires a more rigorous approach to development and maintenance, with a particular emphasis on formal methods to find errors early through rigorous verification and testing, as well as to assist the analysis in locating reported errors.

From the abstract: "The approach is based on automatically learning abstract and succinct protocol models at different layers of the target software stack. Compositional reasoning is then applied to find errors effectively and precisely. The generated models are leveraged to help with error localization and diagnosis."

Kahsai, working this fall with a team of students from the Information Networking Institute's MSIT program, will focus on the Android Bluetooth software stack, developing apps that exercise different functionalities of the Bluetooth Stack and test cases for each function, as well as researching, reproducing and fixing noted bugs.