Today, we see an explosion of Internet-enabled devices, applications, and services. In conjunction, we also observe increasing performance (e.g., network latency, application-perceived quality of experience) and security expectations from users and applications. These requirements have historically outpaced innovation in networking and network security. This creates significant stresses on the network infrastructure, and leads to suboptimal and insecure end-to-end systems. The overarching vision of my research is to provide practical-yet-rigorous solutions that enable network infrastructures to keep up with (and exceed) the evolving performance and security requirements of the applications they enable.
Over the last few years, my research has focused on two key problem domains: (R1) Programmable “software-defined” network functions for delivering better performance and security; and (R2) Network support for improving user-perceived performance of large-scale content distribution applications. A recurring theme of my work in both domains has been to strike a pragmatic balance between architectural rigor and real-world deployability. The research in both domains has rich connections to other domains of computing (e.g., optimization, machine learning, control theory, formal verification).
For more specific projects, please see: https://users.ece.cmu.edu/~vsekar/