The EPOC Project - Energy Proportional and Opportunistic Computing System

Beldiceanu, Nicolas; Feris, Barbara Dumas; Gravey, Philippe; Hasan, Sabbir; Jard, Claude; Ledoux, Thomas; Li, Yunbo; Lime, Didier; Madi-Wamba, Gilles; Menaud, Jean-Marc; Morel, Pascal; Morvan, Michel; Moulinard, Marie-Laure; Orgerie, Anne‐Cécile; Pazat, Jean-Louis; Roux, Olivier; Sharaiha, Ammar

doi:10.5220/0005487403880394

Cited by 8 publications

(6 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Specially at 56 Gb/s using the APD receiver, since the former can only fit in the 100-GHz grid, limiting the maximum number of connected entities to 29, compared to 37, for PAM4 using a 50-GHz grid. As detailed in Beldiceanu et al (2015b), current servers using Solid-State Drives (SSD) can reach 100 Gb/s throughput capacity and thus, it can make efficient use of the optical network designed in the EPOC project.…”

Section: Capacity Issuementioning

confidence: 99%

Towards energy-proportional clouds partially powered by renewable energy

Beldiceanu¹,

Feris²,

Gravey³

et al. 2016

Computing

Self Cite

View full text Add to dashboard Cite

With the emergence of the Future Internet and the dawning of new IT models such as cloud computing, the usage of data centers (DC), and consequently their power consumption, increase dramatically. Besides the ecological impact, the energy consumption is a predominant criterion for DC providers since it determines the daily cost of their infrastructure. As a consequence, power management becomes one of the main challenges for DC infrastructures and more generally for large-scale-distributed systems. In this paper, we present the EpoCloud prototype, from hardware to middleware layers. This prototype aims at optimizing the energy consumption of mono-site Cloud DCs connected to the regular electrical grid and to renewable-energy sources.

show abstract

Section: Capacity Issuementioning

confidence: 99%

Towards energy-proportional clouds partially powered by renewable energy

Beldiceanu¹,

Feris²,

Gravey³

et al. 2016

Computing

Self Cite

View full text Add to dashboard Cite

show abstract

“…Each PM has limited resources (CPU, RAM, Network) and has its own disk storage. We assume that the data center has no centralized storage system (such as a NAS for instance) as described in [14]. The PMs with different capacities and performances may potentially lead to different energy consumption for a given VM.…”

Section: Problem Formulationmentioning

confidence: 99%

“…If gap(t) ≥ 0 (line [8][9][10][11][12][13][14][15][16][17], it means there is extra renewable energy that may be used to switch on n OFF PMs and to run more jobs (opportunistically). Otherwise, it is necessary to switch off m PMs in order to reduce the current energy consumption.…”

Section: B Gapmentioning

confidence: 99%

Opportunistic Scheduling in Clouds Partially Powered by Green Energy

Orgerie

Menaud

2015

2015 IEEE International Conference on Data Science and Data Intensive Systems

Self Cite

View full text Add to dashboard Cite

The fast growth of demand for computing and storage resources in data centers has considerably increased their energy consumption. Improving the utilization of data center resources and integrating renewable energy, such as solar and wind, has been proposed to reduce both the brown energy consumption and carbon footprint of the data centers. In this paper, we propose a novel framework oPportunistic schedulIng broKer infrAstructure (PIKA) to save energy in small mono-site data centers. In order to reduce the brown energy consumption, PIKA integrates resource overcommit techniques that help to minimize the number of powered-on Physical Machines (PMs). On the other hand, PIKA dynamically schedules the jobs and adjusts the number of powered-on PMs to match the variable renewable energy supply. Our simulations with a real-world job workload and solar power traces demonstrate that PIKA saves brown energy consumption by up to 44.9% compared to a typical scheduling algorithm.

show abstract

“…Besides, the EPOC (Energy Proportional and Opportunistic Computing) project [6] consists on energy-aware task execution (from the hardware to application's components) in the context of a small mono-site DC (i.e. less than 50 to 100 servers), connected to the regular electric grid and to local renewable energy sources.…”

Section: Introductionmentioning

confidence: 99%

112-Gbit/s Passive Optical Pod Interconnect for small data centers using Pulse Amplitude Modulation

Feris

Gravey²,

Morel

et al. 2016

2016 International Conference on Optical Network Design and Modeling (ONDM)

Self Cite

View full text Add to dashboard Cite

We compare by numerical simulation several options to increase up to 112 Gbit/s the bit rate used for intradata-center interconnections relying on the recently proposed Passive Optical Pod Interconnect (POPI) architecture, itself derived from the TWIN (Time-Wavelength-Interleaved Network) concept. The motivation for such bit rate is the need to migrate full server content in a small energy-source-aware data center. The analysis focuses on Pulse Amplitude Modulation (4 or 8 PAM), due to its performance/cost ratio. Three types of receivers are considered including a PIN photodiode with integrated semiconductor optical pre-amplifier (SOA-PIN), associated with pre-distortion mechanism. Depending on modulation format and receiver type, the maximum number of connected entities, N, (servers or racks) is limited either by the insertion losses or by the number of channels that can be packed into the band of the fast-tunable-POPI emitters. The highest N value, 48, is obtained using 8 PAM and SOA-PIN in conjunction with a 62.5 GHz grid.

show abstract

The EPOC Project - Energy Proportional and Opportunistic Computing System

Cited by 8 publications

References 9 publications

Towards energy-proportional clouds partially powered by renewable energy

Towards energy-proportional clouds partially powered by renewable energy

Opportunistic Scheduling in Clouds Partially Powered by Green Energy

112-Gbit/s Passive Optical Pod Interconnect for small data centers using Pulse Amplitude Modulation

Contact Info

Product

Resources

About