Validated Thermal Air Management Simulations of Data Centers Using Remote Graphics Processing Units

Sjölund, Johannes; Vesterlund, Mattias; Delbosc, Nicolas; Khan, Amirul; Summers, J. L.

doi:10.1109/iecon.2018.8591192

Cited by 6 publications

(5 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Consequently, server relocation is possible due to the simulator's flexibility. Sjölund et al [36] proposed interactive simulations of data centers on remote GPU utilizing VirtualGL and prioritized the multithreaded implementation of an existing lattice Boltzmann method code. Clement et al [37] followed the Internet of simulation approach to build their simulation.…”

Section: Simulationmentioning

confidence: 99%

Research advances on AI-powered thermal management for data centers

Liu

Aljbri

Song

et al. 2022

Tsinghua Sci. Technol.

View full text Add to dashboard Cite

Given the complex nature of data centers' thermal management, which costs too many resources, processing time, and energy consumption, thermal awareness and thermal management powered by artificial intelligence (AI ) are the targeted study. In addition to a few research on AI techniques and models, other strategies have also been introduced in recent years. Data center models, including cooling, thermal, power, and workload models, and their relationship are factors that need to be understood in the optimal thermal management system.Simulation approaches have been proposed to help validate new models or methods used for scheduling and consolidating processes and virtual machines (VMs), hotspot identification, thermal state estimation, and power usage change. AI-powered thermal optimization leads to improved process scheduling and consolidation of VMs and eliminates the hotspot from happening. At present, research on AI-powered thermal control is still in its infancy. This paper concludes with four issues in thermal management, which will be the scope of further research.

show abstract

Section: Simulationmentioning

confidence: 99%

Research advances on AI-powered thermal management for data centers

Liu

Aljbri

Song

et al. 2022

Tsinghua Sci. Technol.

View full text Add to dashboard Cite

show abstract

“…Future work. As next steps, the authors plan to allow for more varying features, such as changing outside temperature and varying loads, as well as, use models developed by Sjölund et al [12] for training and validating the RL agent. The relative performance of the RL agent and baseline can then be investigated under simulation of a more realistic data center environment, to elicit whether the RL agent adapts better to increased complexity.…”

Section: Summary and Future Workmentioning

confidence: 99%

Towards a Holistic Controller

Heimerson

Brännvall

Sjölund

et al. 2021

Proceedings of the Twelfth ACM International Conference on Future Energy Systems

View full text Add to dashboard Cite

The increased use of cloud and other large scale datacenter IT services and the associated power usage has put the spotlight on more energy-efficient datacenter management. In this paper, a simple model was developed to represent the heat rejection system and energy usage in a small DC setup. The model was then controlled by a reinforcement learning agent that handles both the load balancing of the IT workload, as well as cooling system setpoints. The main contribution is the holistic approach to datacenter control where both facility metrics, IT hardware metric and cloud service logs are used as inputs. The application of reinforcement learning in the proposed holistic setup is feasible and achieves results that outperform standard algorithms. The paper presents both the simplified DC model and the reinforcement learning agent in detail and discusses how this work can be extended towards a richer datacenter model. CCS CONCEPTS• Computer systems organization → Sensors and actuators; • Hardware → Enterprise level and data centers power issues; Temperature control; • Computing methodologies → Reinforcement learning; Modeling and simulation.

show abstract

“…Traditionally, the data center industry has addressed inefficiencies primarily through hardware innovation -more efficient CPUs, better cooling solutions, improved power distribution systems, etc. However, there is huge scope to offer better efficiencies through software-only solutionsexamples include more powerful monitoring and analysis of data center systems [16], predictive capacity planning [17], CFD-based floor layout tools [18], and intelligent scheduling mechanisms [19]. One common factor that is necessary when developing real-world solutions in this space is the need for multi-disciplinary, holistic thinking -it is not enough to view such a system in isolation; instead solutions need to consider the interactions between the physical systems, software systems, business processes, and user behaviors.…”

Section: Takedownmentioning

confidence: 99%

Invited Paper: Improving Data Center Efficiency Through Holistic Scheduling In Kubernetes

Townend¹,

Clement²,

Burdett³

et al. 2019

2019 IEEE International Conference on Service-Oriented System Engineering (SOSE)

View full text Add to dashboard Cite

Data centers are the infrastructure that underpins modern distributed service-oriented systems. They are complex systems-of-systems, with many interacting elements, that consume vast amounts of power. Demand for such facilities is growing rapidly, leading to significant global environmental impact. The data center industry has conducted much research into efficiency improvements, but this has mostly been at the physical infrastructure level. Research into software-based solutions for improving efficiency is greatly needed. However, most current research does not take a holistic view of the data center that considers virtual and physical infrastructures as well as business process. This is crucial if a solution is to be applied in a realistic setting. This paper describes the complex, system-of-systems nature of data centers, and discusses the service models used in the industry. We describe a holistic scheduling system that replaces the default scheduler in the Kubernetes container system, taking into account both software and hardware models. We discuss the initial results of deploying this scheme in a real data center, where power consumption reductions of 10-20% were observed. We show that by introducing hardware modelling into a software-based solution, an intelligent scheduler can make significant improvements in data center efficiency. We conclude by looking at some of the future work that needs to be performed in this area.

show abstract

Validated Thermal Air Management Simulations of Data Centers Using Remote Graphics Processing Units

Cited by 6 publications

References 20 publications

Research advances on AI-powered thermal management for data centers

Research advances on AI-powered thermal management for data centers

Towards a Holistic Controller

Invited Paper: Improving Data Center Efficiency Through Holistic Scheduling In Kubernetes

Contact Info

Product

Resources

About