Manos KapritsosAssistant Professor
Computer Science & Engineering
University of Michigan, Ann Arbor
4824 Bob and Betty Beyster building
I am interested in increasing the reliability of Distributed Systems. Building reliable distributed systems means guarding against both hardware and software failures. My research spans fault-tolerant replication and formal verification of systems software.
In my most recent projects, I have demonstrated how formal verification can be applied to complex distributed systems; I have shown that the ACID paradigm is not fundamentally slow, but can be made to perform well if one separates the ACID guarantees from the mechanism that implements them; and I have shown how to rearchitect the 40-year-old idea of State Machine Replication to accommodate multithreaded execution.
These days, I mostly work on making formal verification practical by applying it to complex systems in new and interesting ways and by automating the process of proving the correctness of these systems.
- July 2019: Our paper I4: Incremental Inference of Inductive Invariants for Verification of Distributed Protocols was accepted at SOSP 2019
- July 2019: Our paper Aegean: Replication beyond the client-server model was accepted at SOSP 2019
- September 2018: I was awarded an NSF grant to rethink fault-tolerant replication beyond the client-server model
- February 2018: I was awarded the Google Faculty Award for 2017
- August 2017: Our paper Vale: Verifying High-performance Cryptographics Assembly Code got the Distinguished Paper Award at USENIX Security 2017
TeachingFall 2019: EECS591 Distributed Systems
Winter 2019: EECS482 Introduction to Operating Systems
Fall 2018: EECS591 Distributed Systems
Winter 2018: EECS591 Distributed Systems
Fall 2017: EECS482 Introduction to Operating Systems
Winter 2017: EECS591 Distributed Systems
Reviewing serviceATC 2019