The News

Visit of an associated professor from Brazil

Détails

Publié le lundi 29 avril 2013 13:39

Martin A. Musicante is associate professor at Universidade Federal do Rio Grande do Norte-UFRN (Natal, Brazil). His main current research topics are XML database theory, Web Services and Programming Languages. Martin A. Musicante will visit us in december to collaborate on new querying approach in the cloud based on map-reduced.

Martin A. Musicante received his B.Sc at ESLAI, Argentina, in 1988 and his M.Sc and PhD in Computer Science at Universidade Federal de Pernambuco, Brazil, in 1989 and 1996, respectively. He is associate professor at Universidade Federal do Rio Grande do Norte-UFRN (Natal, Brazil). He is part of the graduate program in Computer Science at UFRN. He is also a guest researcher at LI - Université François Rabelais Tours, Campus Blois (Blois, France), since 2002 and at LIFO-Université d'Orléans (Orléans, France) since 2008. Martin A. Musicante's main current research topics are XML database theory, Web Services and Programming Languages.

Project: Compiling service coordination queries into map-reduce based programs

The emergence of new architectures like those proposed by Cloud computing open new challenges for querying and retrieving data. The possibility of accessing unlimited resources proposed by the cloud and its associated « pay as U go » model makes it important to change the hypothesis under which data querying and retrieval processes are done. Indeed, instead of conceiving processes and algorithms considering resources availability as a guiding constraint and supposing access to results without any associated constraint, the cloud leads to take into consideration the economic cost of the processes versus the use of computing resources to achieve them, and to present results considering access cost models, and parallel exploitations of available resources.

The three cloud characteristics that we focus on for addressing querying and data retrieval are: (i) the notion of service as unit of construction and access to data and computing functions; (ii) the map – reduce model for the parallel execution of processes ; (iii) coupling SLA and QoS measures for guiding the execution of processes and the consumption of resources.

A new research opportunity emerges in the database domain for studying different map-reduce models for proposing parallel programming strategies for access data taking into consideration the characteristics of the cloud, its economic model and applications quality of service requirements. In this context we aim at proposing approaches based on service coordination techniques for querying and retrieving data on demand. These novel approaches must make evolve existing techniques taking into consideration the resources offered by cloud platforms.