EU commissioned research on energy-efficient data centers with ARM
The project "€ Cloud developed economical server for cloud computing. 2015, first computer with the new design in the data center. ARM Cortex-A9 multicore processors form the technical basis.
The European Commission (EC) has commissioned the chip designer ARM, with the guidance of a project in which it comes to energy-efficient data centers. "Cloud €" aims to develop a very low-power server for Nokia's Ovi service cloud maps, supplies the mobile device with maps for navigation.
ARM is working with Nokia, the specialist in nanotechnology IMEC, Ecole Polytechnique Federale de Lausanne and the University of Cyprus. Euro Cloud has already received an undisclosed grant from the European Commission, should be enough for three years.
The project is based on low-power ARM Cortex-A9 multicore processors and dell vostro 1510 battery 3D-DRAM memory modules. These techniques should reduce the energy consumption.
The € Cloud project members assume that servers are available with the design already developed by them in 2015 in the data center. The aim of the project is to accommodate hundreds of processor cores in a single server. "We must achieve energy-efficient multi-core processors for a breakthrough," said Panagiotis Tsarchopoulos, who is in charge of the European Commission for the project. "Brings € Cloud is an interesting research project, leading European low-power technology in the field of embedded computing in general computer technology."
ARM stated that the company had so far focused on energy savings in mobile devices. Now would you turn to the power consumption in data centers. "The performance and the new features of our recent processor models have reached a level where it makes sense to explore how ARM technology, the major challenges in energy consumption of wired applications - solve, for example in data centers -" said Krisztian Flautner, Vice President of Research at ARM.
Project participants will present their concepts for a low-power server on the € Cloud Cloud homepage. It says: "Cloud computing applications require greater parallelism to thread-level or memory. A parallelization at the level of individual instructions is little. Each client request is answered by a single or multiple independent threads, all on a separate processor to run. "