Grant: Canada Research Chairs

Computer networks have grown into a critical infrastructure in modern society. The way we work, bank, shop, learn, access media, and interact with each other largely relies on computer networks to transfer data correctly and efficiently between computers such as our personal devices and those hosting online services. As such, it is crucial for modern computer networks to operate in a robust, reliable, and efficient manner. Unfortunately, network malfunctions are not uncommon today. Modern networks are extremely complex ecosystems with thousands of sophisticated and constantly evolving software and hardware components. These components closely interact with each other to perform millions of concurrent data transfers. The human engineers and operators who manage this complexity can often make mistakes. In fact, human error -- e.g., errors in configuration updates as opposed to, say, equipment failure -- has been identified as the leading cause of network malfunctions. A malfunction in even one component can have far-reaching and severe consequences; it can cause ripple effects throughout the network, result in failed or agonizingly slow data transfers, and create outages that disrupt millions of lives. The overarching goal of Dr. Arashloo's research is to minimize human error in operating modern computer networks. Dr. Arashloo and her team develop novel techniques and systems that draw from various theoretical and practical areas of computer science to (1) conduct automated formal analysis of network components before deployment to point out potential mistakes, (2) generate correct and efficient implementations for deployment, and (3) monitor the network after deployment to report potential violations of expected behavior in real time. These three strategies have mostly been studied separately, and only for certain families of network components and basic properties (e.g., data from Alice's phone should eventually reach Bob's). Dr. Arashloo's research will bring all three together in cohesive frameworks that can ensure the advanced properties required of modern networks (e.g., data transfers for video calls should not experience high delay). Dr. Arashloo's research will equip network engineers and operators with much-needed rigorous and automated assistance in detecting and preventing errors. As such, it will play a crucial role in making networks robust, reliable, and efficient in their service to the society that is increasingly relying on them.