Observatorio de I+D+i UPM

Memorias de investigación
Managing Consistency for Big Data Applications: Tradeoff s and Self-Adaptiveness
Research Areas
  • Computer systems
In the era of Big Data, data-intensive applications handle extremely large volumes of data while requiring fast processing times. A large number of such applications run in the cloud in order to benefit from cloud features. In this context, replication is essential in order to deal with Big Data challenges. However, replication introduces the major issue of data consistency across different copies. Consistency management is a critical matter for Big Data systems. Strong consistency models introduce serious limitations to systems scalability and performance due to the required synchronization efforts. In contrast, weak and eventual consistency models reduce the performance overhead and enable high levels of availability. However, these models may tolerate, under certain scenarios, too much temporal inconsistency. In this Ph.D thesis, we address this issue of consistency tradeoffs in Big Data systems and applications. First, we focus on consistency management at the storage system level. We propose an automated self-adaptive model that scale up/ down the consistency level at runtime when needed in order to provide as high performance as possible while preserving the application consistency requirements. In addition, we address the consistency management impact on monetary cost in the cloud. Accordingly, we propose a cost-efficient tuning of consistency levels in the cloud. In a third contribution, we study the consistency management impact on energy consumption within the data center. Thereafter, we investigate adaptive configurations of the storage system cluster that save energy. In order to complete our system-side study, we focus on the application level. We propose a behavior modeling approach that apprehends the consistency requirements of the application. Based on the model, we propose an online prediction approach that adapts to the application specific needs at runtime and provides customized consistency. Extensive evaluations on the Grid?5000 testbed and Amazon EC2 demonstrate the efficiency of the introduced approaches.
Mark Rating
  • Director: Maria de los Santos Perez Hernandez (UPM)
  • Autor: Houssem Eddine Chihoub (Université Européenne de Bretagne)
Research Group, Departaments and Institutes related
  • Creador: Grupo de Investigación: Ontology Engineering Group
  • Departamento: Arquitectura y Tecnología de Sistemas Informáticos
S2i 2020 Observatorio de investigación @ UPM con la colaboración del Consejo Social UPM
Cofinanciación del MINECO en el marco del Programa INNCIDE 2011 (OTR-2011-0236)
Cofinanciación del MINECO en el marco del Programa INNPACTO (IPT-020000-2010-22)