Speaker

Nate Foster

Nate Foster

Professor of computer science at Cornell University, visiting researcher at Jane Street, and a founding contributor to two influential network programming languages: NetKAT and P4. His research sits at the intersection of programming language theory and computer networking — a combination he arrived at through a deliberate mid-career leap, leaving a faculty position at Cornell to spend a postdoc at Princeton with Jennifer Rexford and David Walker, then a sabbatical at Barefoot Networks (the programmable-switch startup later acquired by Intel).

His earlier academic work, completed under Benjamin Pierce at the University of Pennsylvania, was on lenses — abstractions for synchronising data between two representations that later became widely used, in generalised form, in the Haskell programming community. That background in formal language theory proved unexpectedly portable: NetKAT’s design was discovered to align precisely with Kleene algebra with tests (KAT), a decades-old mathematical framework, giving network packet-forwarding the same kind of rigorous semantic foundation that type theory provides for programs.

At Jane Street, Foster has helped design Butane, a system that compiles high-level network policy into BGP configurations for the firm’s global wide-area network. The system gives engineers a code-reviewed, formally validated alternative to editing individual router configurations by hand, and includes tooling for ‘what-if’ analysis and capacity planning against a semantic model of BGP.

Core positions

Foster argues that a computer network is a program — one that should be specified, composed, tested, and verified using the same methods applied to software, rather than configured device by device through vendor interfaces and standards-body processes. He holds that formal methods derive their greatest value not from catching individual bugs but from enabling engineers to take larger, faster steps by making the consequences of proposed changes legible before deployment. On research culture, he is a vocal advocate for honest retrospection about ideas that did not prevail: the networking community’s willingness to examine failures publicly is, in his view, what keeps it innovative.

In the wiki