Effect of an Incentive Implementation for Specifying Accurate Walltime in Job Scheduling

Stolf

2022 IEEE 34th International Symposium on Computer Architecture and High Performance Computing (SBAC-PAD)

2022

Both academics and industry have engaged their efforts in reducing greenhouse gas (GHG) emissions of Information and Communications Technology (ICT). Data centers are one of the most electricity-expensive ICT actors due to their uninterrupted service. Reducing the usage of brown energy or migrating to green energy using renewable sources (RES) is a way to reduce these emissions. However, this migration is not straightforward due to intermittence from these sources. This work is part of Datazero 2 ANR project. This project aims to design a data center powered only by RES production and storage elements. This architecture requires several decisions at different levels of management. This project divides the problem into two groups: offline and online. On the offline side, it uses renewable and workload predictions to prepare an offline plan with a power envelope (power delivered to the servers) and hints on how to manage the storage (batteries and hydrogen). Given the offline plan, the article's contribution is how to deal with real and dynamic power constraints online while keeping the planned storage level at the end. So, this article proposes policies to modify the plan according to the changes in predictions. We evaluate these policies in a homogeneous and heterogeneous data center. The results demonstrate that our policies could approach the storage level and improve Quality of Service (QoS) independently of data center infrastructures.

Section: B Online Model 1) Schedulingmentioning

confidence: 99%

Analyzing Power Decisions in Data Center Powered by Renewable Sources

Stolf

2022 IEEE 34th International Symposium on Computer Architecture and High Performance Computing (SBAC-PAD)

2022

“…After defining the server configuration, the scheduler must assign jobs to the servers. This article focuses mainly on CPUbound scientific high-performance computing (HPC) jobs [21], [22]. In this application class, the user executes one application and waits for its end [23].…”

Section: Problem Statementmentioning

confidence: 99%

“…In this application class, the user executes one application and waits for its end [23]. Another characteristic of this job type is that the schedule can delay the processing start [22]. This feature allows choosing the best moment to run it.…”

Section: Problem Statementmentioning

confidence: 99%

“…Since the server configuration changes the server's state and speed, the scheduling must consider these modifications to define the placement. The scientific HPC applications are submitted with the following characteristics: arrival date, the number of parallel resources, and walltime defined by the user [21], [22]. The first property gives the job submission time.…”

Section: Problem Statementmentioning

confidence: 99%

“…So, it must infer it using the execution time, for example. The execution time could also be estimated by the walltime [22]. The main objective of the scheduling is to find the first possible moment after Ar j to place job j in P l j servers, meeting the constraints from Equations 4 and 5.…”

Section: S S=0mentioning

confidence: 99%

See 2 more Smart Citations

Evaluation of Heuristics to Manage a Data Center Under Power Constraints

Stolf

2022 IEEE 13th International Green and Sustainable Computing Conference (IGSC)

2022

HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L'archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d'enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Mixing Offline and Online Electrical Decisions in Data Centers Powered by Renewable Sources

Stolf²,

IECON 2022 – 48th Annual Conference of the IEEE Industrial Electronics Society

2022

Since the Paris Agreement, academics and industry have dedicated efforts to reducing Information and Communications Technology (ICT) greenhouse gas (GHG) emissions. Data centers play a key role in ICT electricity consumption since they are built to run 24/7. One way to reduce these emissions is to switch from brown energy to green energy from renewable sources (RES). However, RES introduce several uncertainties due to their intermittence. This work is part of the Datazero2 Project. This project aims to design a data center powered only by RES production and storage elements. Datazero2's clean-bydesign data center requires several decisions at different levels of management. Usually, research works focus on offline decisions (e.g., decisions for the next three days) or online decisions (e.g., decisions in real-time). This work proposes a mix of offline and online decisions producing a reliable and change-aware solution. While the offline focuses on the long-term renewable usage plan, online adapts it according to the power fluctuations. The results demonstrate that this mixed management approximates the planned and real storage levels with a low impact on the Quality of Service (QoS).