Vasileios (Vasilis) Giotsas is joining the Resilient Networks Lab (ResNetLab) as a Research Engineer. He comes to PL from Lancaster University, where he was an Assistant Professor, after stints at Technischen Universität Berlin and the University of California San Diego. His interests include network measurement and telemetry, topology mapping, and routing.
We asked Vasilis about how we got here and what the future looks like:
What path brought you to PL?
I have done extensive research in measuring and analyzing the Internet ecosystem, through which it became evident to me that one of the most apparent shortcomings of the current network is the increasing consolidation of services and infrastructure into a few hypergiant service providers. This centralization not only makes the Internet more vulnerable to targeted attacks and censorship, but also creates problematic network economics. For example, the cost of conducting a massive DDoS attack is only a few hundred dollars, while the cost of mitigating such attacks is on the order of tens of million dollars. My research in the design of more resilient Internet protocols introduced me to IPFS. After watching Juan Benet’s talk at Stanford University about the decentralized web I realized that Protocol Labs has embarked on a very special mission to evolve the web and realize some of the most ambitious academic ideas in peer-to-peer networking. The open and transparent approach in the development of IPFS and the rest of Protocol Labs' stack has made it possible to create an enthusiastic community around the projects, which has led to a vibrant ecosystem of decentralized applications. All of this made the opportunity to work with Protocol Labs a very exciting prospect.
What are you working on?
I am working with the Resilient Networks Lab, which is part of Protocol Labs Research. My aim is to improve the observability of public, user-run, permissionless p2p networks, which have unpredictable structure and dynamics. We believe that improving the transparency of public p2p networks and illuminating macroscopic aspects of their behavior are necessary in order to promote a healthy ecosystem of protocol developers where hypotheses are tested and verified through real-world measurements. Having the ability to get information from the real-time performance of the network can help tremendously in both real-time decision-making and in long-term protocol design. To succeed in this effort, I will coordinate and get input from PL designers and engineers, and the broader open-source community focused on decentralized p2p applications.
What future technology are you most excited about?
I am really excited about Fully Homomorphic Encryption (FHE), which will allow arbitrary computations on encrypted data, and upon decryption the results will be correct. FHE can pave the way to decentralized peer-to-peer computing since it will alleviate privacy concerns with outsourcing the management of sensitive data to third parties. While there has been a lot of progress in constructing and improving FHE schemes, performance is the main barrier for adopting FHE in real-world applications. Somewhat Homomorphic Encryption (SHE) can be a viable alternative for many applications. SHE supports a limited set of homomorphic operations but can have much better performance than FHE while still being useful to many applications currently deployed on centralized cloud computing systems, such as the processing of medical information.