TAO THANH Vinh
Supervision : Marc SHAPIRO
Co-supervision : RANCUREL Vianney
Ensuring Availability and Managing Consistency in Geo-Replicated File Systems
Geo-distributed systems suffer from high latency and network partitions. Because of this, and to ensure high availability, such systems typically commit updates locally, with no latency, and propagate them in the background. Such optimistic replication faces two major challenges: (i) detecting conflicts between concurrent updates and resolving them in a way meaningful for users, while maintaining system integrity invariants; and (ii) supporting legacy applications that are not prepared to deal with concurrency anomalies.
Our PhD research addresses these challenges for the specific use case of a large-scale geo-distributed file system. This is a good showcase: indeed, a geo-distributed file system has a complex hierarchical structure; maintaining the file-system invariants (e.g., tree structure) against parallel updates is challenging; and applications view the file system through the legacy POSIX API.
We present our solution to the above problems in the design and implementation of a prototype geo-distributed file system, named Tofu. Its design includes a new session abstraction to support the legacy API, while allowing optimistic updates. Unlike previous approaches, our solution is based on a formal model covering all aspects of a Unix-like file system, including directories, inodes, hard links, etc. It is able to detect all conflicts on those data structures, and resolves them in a way that we believe users will find generally reasonable. Experiments show that Tofu is highly scalable, and incurs linear overhead, improving over existing academic and industrial systems.
Defence : 12/08/2017 - 10h - Site Jussieu 25-26/105
Jury members :
Mme. BOUCHENAK Sara, INSA de Lyon [Rapporteur]
M. MOLLI Pascal, Université de Nantes [Rapporteur]
Mme. BIENIUSA Annette, University of Kaiserslautern
Mme. LAWALL Julia, Inria de Paris
M. QUEMA Vivien, Grenoble INP
M. RANCUREL Vianney, Scality
M. SHAPIRO Marc, Inria de Paris
- V. Tao Thanh : “Ensuring Availability and Managing Consistency in Geo-Replicated File Systems”, thesis, defence 12/08/2017, supervision Shapiro, Marc, rapporteurs : RANCUREL Vianney (2017)
- V. Tao Thanh, M. Shapiro, V. Rancurel : “Merging Semantics for Conflict Updates in Geo-Distributed File Systems”, ACM Int. Systems and Storage Conf. (Systor), Haifa, Israel, pp. 10.1-10.12 (2015)
- M. Segura, V. Rancurel, V. Tao Thanh, M. Shapiro : “Scality’s experience with a geo-distributed file system”, Middleware 2014, Bordeaux, France, pp. 31-32, (ACM) (2014)