Team-project REGAL

Principled approaches to eventual consistency

Marc Shapiro — 2012-07-15T17:00:57Z

Data replication is a fundamental mechanism of distributed systems, for performance, availability and fault tolerance. To sidestep the incompatibility between strong consistency and fault-tolerance, an attractive approach is eventual consistency (EC). EC allows a replica to update without synchronising ; concurrent updates are reconciled in the background. However, existing EC algorithms are often ad-hoc and/or unpredictable ; conflict resolution requires synchronisation and is complex and (...) - Post-Docs

Principled approaches to eventual consistency

Marc Shapiro — 2012-07-15T16:59:04Z

Data replication is a fundamental mechanism of distributed systems, for performance, availability and fault tolerance. To sidestep the incompatibility between strong consistency and fault-tolerance, an attractive approach is eventual consistency (EC). EC allows a replica to update without synchronising; concurrent updates are reconciled in the background. However, existing EC algorithms are often ad-hoc and/or unpredictable; conflict resolution requires synchronisation and is complex and (...) - Post-Doc

Highly Scalable Data Store

Masoud Saeida Ardekani — 2011-11-22T19:12:07Z

Highly scalable data stores are fundamental for today Internet services, e.g., social networking sites. Modern data stores replicate data for both performance and availability. Full replication has limited scalability as every replica must replay every update. Partial replication [1, 2] addresses this problem by replicating only a subset of available data at each replica. This allows the system to handle independent parts of the workload in parallel, and has the potential to scale better. (...) - Stages

Highly Scalable Data Store

Masoud Saeida Ardekani — 2011-11-22T19:03:01Z

Performance models for concurrent programs

Annette Bieniusa — 2011-11-22T08:24:08Z

There exists an abundance of synchronization protocols and algorithms for concurrent programming : mutual exclusion schemes with locks and barriers, optimistic executions with transactions, explicit data exchange via message passing, and many more. They all provide some means to achieve data synchronization between different threads, but which one is best ?
A comparison of these synchronization schemes can be done along several lines. In theory, complexity measures usually are based on (...) - Stages

Performance models for concurrent programs

Annette Bieniusa — 2011-11-14T16:32:59Z

There exists an abundance of synchronization protocols and algorithms for concurrent programming: mutual exclusion schemes with locks and barriers, optimistic executions with transactions, explicit data exchange via message passing, and many more. They all provide some means to achieve data synchronization between different threads, but which one is best?
A comparison of these synchronization schemes can be done along several lines. In theory, complexity measures usually are based on (...) - Internships

Thesis position at LIP6-REGAL/Telecom Bretagne.

Monnet Sébastien — 2011-08-04T13:36:55Z

Thesis position at LIP6-REGAL/Telecom Bretagne.
Thesis subject : Placement and replication strategies for next generation on-line data storage
Context
This thesis is funded by the ODISEA FUI project, which gathers five partners : Orange, University Pierre-et-Marie-Curie (the LIP6-REGAL team), UbiStorage, Technicolor, Institut Telecom (including Telecom Bretagne).
ODISEA aims at designing new on-line data storage and data sharing solutions. Current solutions rely on big data centers, which (...) - Thèses

Website

Pierre Sens — 2011-01-17T12:15:54Z

http://coccinelle.lip6.fr/ - Coccinelle

Website

Pierre Sens — 2011-01-17T09:00:39Z

http://coccinelle.lip6.fr/ - Coccinelle

Website

Pierre Sens — 2011-01-17T08:51:37Z

https://gforge.inria.fr/projects/pastis/ - PASTIS