Papers Papers/2020



Paper Unlock

Authors: Giovanni Agosta 1 ; Alessandro Barenghi 2 ; Tomasz Ciesielczyk 3 ; Rahul Dutta 4 ; William Fornaciari 1 ; Thierry Goubier 5 ; Jens Hagemeyer 6 ; Lars Kosmann 7 ; Nicholas Mainardi 1 ; Ariel Oleksiak 3 ; Gerardo Pelosi 2 ; Wojciech Piatek 3 ; Christian Pieper 8 ; Mario Porrmann 6 ; Daniel Schlitt 8 and Michele Zanella 1

Affiliations: 1 DEIB, Politecnico di Milano, Italy ; 2 Politecnico di Milano, Italy ; 3 Poznan Supercomputing and Networking Center, Poland ; 4 Huawei Technologies GmbH, Germany ; 5 Institut List, CEA, Paris-Saclay University, France ; 6 CITEC, Bielefeld University, Germany ; 7 CEWE Stifung & Co. KGaA, Germany ; 8 OFFIS e. V., Germany

Abstract: The H2020 project Modular Microserver DataCentre (M2DC) investigates, develops and demonstrates a modular, highly-efficient, cost-optimized server architecture composed of heterogeneous microserver computing resources. The M2DC architecture can be tailored to meet requirements from various application domains such as image processing, cloud computing or HPC. To achieve this, M2DC is built on three pillars. (1) The RECSjBox, a flexible server architecture fully configurable with respect to application requirements supports the full range of microserver technologies, including low power ARM processors or FPGA accelerators as well as high performance x86 or GPU devices. (2) Advanced management strategies as well as system efficiency enhancements (SEE) improve the behaviour of the system during runtime, thereby addressing application acceleration, communications and monitoring & management. Moreover, an intelligent management module complements the middleware by proactive workload, therm al and power management to increase the energy efficiency. (3) Welldefined interfaces to the software ecosystem enable easy integration of the customized RECSjBox system into the existing data centre landscape. By integrating into OpenStack for bare metal orchestration of the microservers, the applicability in today’s data centre is granted. Current project results include new microserver designs based on ARM64 and Intel Stratix 10. The document presents TCO estimations and baseline benchmarks to show the high potential of accelerators for the targeted applications including image processing, Internet-of-things (IoT) data processing and others. (More)


Sign In Guest: Register as new SciTePress user now for free.

Sign In SciTePress user: please login.

PDF ImageMy Papers

You are not signed in, therefore limits apply to your IP address

In the current month:
Recent papers: 100 available of 100 total
2+ years older papers: 200 available of 200 total

Paper citation in several formats:
Agosta, G.; Barenghi, A.; Ciesielczyk, T.; Dutta, R.; Fornaciari, W.; Goubier, T.; Hagemeyer, J.; Kosmann, L.; Mainardi, N.; Oleksiak, A.; Pelosi, G.; Piatek, W.; Pieper, C.; Porrmann, M.; Schlitt, D. and Zanella, M. (2017). The M2DC Approach towards Resource-efficient Computing. In OPPORTUNITIES AND CHALLENGES for European Projects - EPS Portugal 2017/2018, ISBN 978-989-758-361-2, pages 150-176. DOI: 10.5220/0008862601500176

@conference{eps portugal 2017/201817,
author={Giovanni Agosta. and Alessandro Barenghi. and Tomasz Ciesielczyk. and Rahul Dutta. and William Fornaciari. and Thierry Goubier. and Jens Hagemeyer. and Lars Kosmann. and Nicholas Mainardi. and Ariel Oleksiak. and Gerardo Pelosi. and Wojciech Piatek. and Christian Pieper. and Mario Porrmann. and Daniel Schlitt. and Michele Zanella.},
title={The M2DC Approach towards Resource-efficient Computing},
booktitle={OPPORTUNITIES AND CHALLENGES for European Projects - EPS Portugal 2017/2018,},


JO - OPPORTUNITIES AND CHALLENGES for European Projects - EPS Portugal 2017/2018,
TI - The M2DC Approach towards Resource-efficient Computing
SN - 978-989-758-361-2
AU - Agosta, G.
AU - Barenghi, A.
AU - Ciesielczyk, T.
AU - Dutta, R.
AU - Fornaciari, W.
AU - Goubier, T.
AU - Hagemeyer, J.
AU - Kosmann, L.
AU - Mainardi, N.
AU - Oleksiak, A.
AU - Pelosi, G.
AU - Piatek, W.
AU - Pieper, C.
AU - Porrmann, M.
AU - Schlitt, D.
AU - Zanella, M.
PY - 2017
SP - 150
EP - 176
DO - 10.5220/0008862601500176