Regal is a joint research group of LIP6 and INRIA-Rocquencourt. Regal investigates large-scale distributed systems, and especially P2P architectures. An important focus is large-scale replication in very dynamic settings. We investigate adaptive algorithms, in order to react to changes in the environment and in the application. Here are some of our research areas:
Data management in large scale configuration: to deploy and locate data, and to manage consistency.
System monitoring and failure detection: we investigage failure detectors with provable properties in dynamic environments.
Replication: replication of data improves availability and responsiveness but updates raise the issue of consistency. We have several interests in this area: fault-tolerant replication techniques, optimistic techniques (which allow local updates but cause replication divergence) and adaptive replication. Our research aims to compare and evaluate existing algorithms and to combine their best features into new, high-performance protocols.
Dynamically-adaptative operating systems
Team REGAL merges the "Operating Systems" group of LIP6, and the SOR project of INRIA. Marc Shapiro was the leader of SOR until 1998, then by Mesaac Makpangou. You can view the list of contributors here. Here is a list of some of the projects of the two groups prior to their merger:
Fault tolerance in local area network (STAR)
Unification of fault tolerance and load balancing (GATOSTAR)
Distributed shared memory ( Perdis ), (CLRC);
Cooperative Web cache( Relais );
Management of the distributed objects ( SSPC , Cadmium );
System of configuration and dimensionment of the systems of masks of company ( Saperlipopette! );
Annotation services ( Pharos );
Adaptable replicated bjects ( BOAR ).
Flexible binding system ( Hobbes );
Flexible storage system( BOSS ); - Garbage collectors in a persistent distributed shared memory ( Larchant );
Generic and configurable logging system ( Kitlog );
Flexible object-oriented Operating System ( SOS );