White Rose University Consortium logo
University of Leeds logo University of Sheffield logo York University logo

Holistic and Energy-Efficient Management of Datacentres

Al-Anii, Yaser Talal Saber (2017) Holistic and Energy-Efficient Management of Datacentres. PhD thesis, University of Leeds.

Al-Anii_Yaser T S_ School of Mechanical Engineering_University of Leeds_PhD_2017.pdf - Final eThesis - complete (pdf)
Available under License Creative Commons Attribution-Noncommercial-Share Alike 2.0 UK: England & Wales.

Download (8Mb) | Preview


The overall power consumption of datacentres is increasing tremendously due to the high demand of digital services. Moreover, the cooling load contributes up to 50% of the power consumption due to the higher densities of newer versions of servers. However, there is an increased awareness in the operations of the sub-systems, i.e. workload, cooling load and power consumption. This awareness of the interactions between the sub-systems provides a better understanding for maintaining the datacentre as an energy-efficient infrastructure. A direct contact liquid cooling technology is examined extensively by retrofitting to an air-cooled server. First the conventional SunFire V20z air-cooled server is benchmarked via SPECpower_ssj2008 workload to obtain some standard values. The server is placed inside a wind tunnel to ensure a controllable environment. Then an overall evaluation of the retrofitted server is presented and compared with the standard server. The retrofitted server shows a reduced cooling power consumption of 29%. In addition, the performance to power ratio increases by 10% comparing to the conventional server. The liquid cooling technology keeps the central processes units (CPUs) up to 10 oC colder than the air-cooled server. Furthermore, the new server operates in an 88% lower noise after the replacement of four fans by two pumps. However, the main restriction of using such a solution is the risk of bringing water into the microelectronics due to leakage and condensation of water. A fully immersed encapsulated server is then investigated to assess the validity of simulating the immersed server as a porous layer. This simulation uses Darcy flow with mass, momentum and energy conservation equations. The model shows a quantitive and qualitative accuracy compared to the previous work. The model shows that the distance between processors has a strong effect on the thermal behaviour of the encapsulated server by 13.3% compared to servers’ dimensions. Moreover, the model presents the optimal design and geometry of an encapsulated server with respect to the thermal performance. Although the model is simple, it can be used for an initial prediction of the server design. This is due to the limitation of capturing the thermal behaviour of a full model. A holistic power consumption model is presented to capture the interactive relationships between servers’ sub-system. The power model relies on experimental work and is constructed based on the collected data from different cooling configurations. The model captures a detailed breakdown of the power consumption and therefore presents an accurate calculation of the partial power usage effectiveness metric. The results are limited to one microelectronic architecture within a specific IT load type. However, the results show that reducing the cooling load by 7% and increasing the performance by 5% leads to lower the partial power usage effectiveness by 1%. Finally, the current study explores the usage of an evaporative air handling unit for energy-efficient datacentres. The air handling unit is capable of run dry and wet cooling operation. The cooling system operated successfully during July and August 2016, in Leeds. The wet cooling has a higher thermal performance than the dry cooler due to the large heat capacity of water compared to air. Therefore, the wet cooling configuration records a power usage effectiveness lower than the dry cooling by about 6.4%.

Item Type: Thesis (PhD)
Related URLs:
Keywords: Datacentre, Liquid Cooling, Holistic Management, Unified Power Model, Immersed Server.
Academic Units: The University of Leeds > Faculty of Engineering (Leeds) > School of Mechanical Engineering (Leeds)
Identification Number/EthosID: uk.bl.ethos.713234
Depositing User: Yaser Talal Saber Al-Anii
Date Deposited: 15 May 2017 09:21
Last Modified: 18 Feb 2020 12:48
URI: http://etheses.whiterose.ac.uk/id/eprint/16723

You do not need to contact us to get a copy of this thesis. Please use the 'Download' link(s) above to get a copy.
You can contact us about this thesis. If you need to make a general enquiry, please see the Contact us page.

Actions (repository staff only: login required)