Title: Decomposing consistency: It’s the invariants, stupid
By: Marc Shapiro (INRIA & Sorbonne Universites)
Abstract:
Maintaining consistency is a central issue of distributed systems. However, there is a fundamental design tension between highly-synchronised models, which are easy to use, and ones with less synchronisation but weaker properties. This results in a complex and confusing design space.
How do consistency models differ, and which one is right for me? To answer, we describe them in terms of elementary, orthogonal properties. We show (i) what are the guarantees provided by a given property (i.e., what kind of application invariants that it ensures), and, (ii) dually, the opportunities for parallelism and implementation freedoms that it provides. Then, we analyse some common classes of application invariants, and show their relation to the above properties. Finally, we present our current work on a top-down invariant-directed approach to design a protocol that has just enough synchronisation to ensure the application’s invariants.
Short bio:
Marc Shapiro does his research on distributed computer systems, data replication and consistency algorithms, and distributed garbage collection. He invented the proxy concept, which is now universal on the Internet. He published at SOSP and OSDI, the two most prestigious venues of the area (one of the only two French papers at both venues). He was instrumental in the creation of EuroSys, the main European venue in the area. He authored 64 international publications, 17 recognised software systems, and four patents. Dr Shapiro’s research started with a PhD from Université Paul Sabatier for research performed at at LAAS in Toulouse, France, followed by a post-doc at MIT, and a researcher position at CMIRH. He is a researcher at INRIA since 1984. He spent a one-year sabbatical at Cornell, and he led the Cambridge Distributed Systems group at Microsoft Research Cambridge. He is currently a Senior Researcher for INRIA Paris-Rocquencourt, in the Regal group, located at LIP6.