Research area
Cloud Computing
Abstract
The goal of this activity is to provide solutions for effective, efficient, elastic resource management across multiple clouds for a wide range of application types (e.g., applications that fit the MapReduce paradigm and data-intensive applications) in federated public and private cloud infrastructures as (an) extension(s) of the current systems of the partners. Even though private and, in particular, public clouds are often based on very large sets of resources (e.g., in large data centers), they have to be used efficiently because 1) even in clouds there may still be resource contention as public cloud owners may want to accommodate as many customers as possible, and because 2) users of public clouds want to select the optimal resources from cloud federations for their applications while spending as little money as possible for their cloud usage. The innovation pursued by this activity proposal in year 1 (2012) is to create software demonstrators based on the research performed and the software already created by the partners in carrier projects that enable effective, efficient, and elastic use of cloud federations. This outcome is novel in that currently, no such universal tools/demonstrators for elastic cloud computing exist. As the current software systems for resource management of the partners in this activity have overlapping objectives and we are planning this activity for more than one year, two important additional goals in year 1 are to investigate the similarities and complementarities of the resource management solutions of the different partners in the activity in order to integrate the software and testbeds or to make them inter-operate in later yeasr, and to explore more collaboration with industry.
Partners
Institut Telecom (FR), INRIA (FR), KTH/SICS (SE), TU/e (NL)
Time frame
2012+
Funding
350+ kEURO, EU EIT
People involved
Dick Epema, Associate Professor