PaPoC 2024 will take place on Monday, April 22nd, 2024. It is an in-person only event, happening in conjunction with the EuroSys 2024 conference. The workshop will be located at the Royal Olympic Hotel. To see how to reach the conference area, please follow this link.

Proceedings

Proceedings in the ACM Digital Library.

Registration: 08:30 to 09:00

Opening & Session 1 (Distributed Systems): 09:00 to 10:30

Abstract
Traditional strong consistency maintains the same total order of operations across replicas. The total order is the immediate source of desirable properties such as integrity, convergence and recency. However, maintaining the total order has proven to inhibit availability, responsiveness and scalability. Weaker notions uphold performance; however, they forfeit the total order, and hence its favorable properties. Thus, application programmers face non-trivial choices between a spectrum of consistency notions. In this talk, I present a project that, given a sequential class with the declaration of its integrity and recency requirements, automatically synthesizes a correct-by-construction replicated class that guarantees integrity, convergence and recency, and avoids coordination as much as possible. The approach is based on a novel sufficient condition called well-coordination for integrity and convergence which requires certain orders between conflicting and dependent operations. Our synthesis tools apply automatic solvers to statically derive the conflict and dependency relations. They then reduce the coordination avoidance problem to classical graph minimization problems, and use the results to instantiate parametric coordination protocols, and synthesize replicated stores. Further, in order to preserve recency specifications, replicas use solvers at runtime to bound the set of pending operations. We applied this approach to message-passing networks and then extended it to RDMA networks. These results lead to a sequence of presentations at POPL ’19, CAV ’20 and PLDI ’22. At the end, I briefly talk about our new thrusts to accelerate replicated stores on FPGAs, and replicated virtual objects on AR devices.

Bio
Mohsen Lesani is an associate professor at the Computer Science and Engineering department of the University of California, Santa Cruz. He spent his postdoc at MIT and obtained his PhD from UCLA. His research interests are reliability and security of software systems especially concurrent and distributed systems. He received the NSF CAREER award in 2020 and the DARPA YFA in 2022. His research has been recognized as SIGPLAN Research Highlight in 2019 and received the distinguished paper award at OOPSLA ‘18.

Coffee Break: 10:30 to 11:00

Session 2 (Coordination-Free Systems): 11:00 to 12:30

Chair: Philipp Haller

Lunch: 12:30 to 13:45

Abstract
Distributed storage systems are the foundation of large-scale applications; they provide scalable data access to application code written by frontend programmers. At the heart of their design is a tension between guarantees and performance. Their guarantees—e.g., their consistency model—have a major effect on the correctness, simplicity, and development speed of application code. Their performance—e.g., their latency—has a major effect on the performance of the application because they are accessed many times during an application-level event like a page load.

This talk covers a line of research on building better distributed systems by discovering the principles that characterize the tension between guarantees and performance. I will describe several new principles: from defining optimality and proving impossibility results to introducing new consistency models. And I will show how we can use our new understanding to build better systems that ultimately deliver better performance for applications without sacrificing guarantees.

Bio
Wyatt Lloyd is an Associate Professor of Computer Science at Princeton University. His research focuses on the distributed systems that underlie the architecture of large-scale Web sites, cloud computing, and big data. He received his Ph.D. from Princeton University in 2013, spent the next year as a Postdoctoral Researcher at Facebook, and spent 3 years as an Assistant Professor of Computer Science at the University of Southern California before returning to Princeton in 2017. He won the NSF CAREER award and is a Sloan Fellow.

Session 3 (Verifying Databases): 14:45 to 15:30

Chair: Philipp Haller

Coffee Break: 15:35 to 16:00

Session 4 (Byzantine Systems) and Closing: 16:00 to 17:20

Chair: Natacha Crooks