2020
DOI: 10.1126/science.aba3758
|View full text |Cite
|
Sign up to set email alerts
|

Recalibrating global data center energy-use estimates

Abstract: Growth in energy use has slowed owing to efficiency gains that smart policies can help maintain in the near term

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

2
350
0
17

Year Published

2020
2020
2023
2023

Publication Types

Select...
6
3

Relationship

0
9

Authors

Journals

citations
Cited by 726 publications
(369 citation statements)
references
References 9 publications
2
350
0
17
Order By: Relevance
“…1). However, the actual development has turned out very different, according to [6] or [7] (Fig. 1).…”
Section: Kupzog Friederich Ait Austrian Institute Of Technology Gmbmentioning
confidence: 99%
“…1). However, the actual development has turned out very different, according to [6] or [7] (Fig. 1).…”
Section: Kupzog Friederich Ait Austrian Institute Of Technology Gmbmentioning
confidence: 99%
“…The objective of the MILP model is to minimize the total network power consumption (TNPC), total compute power consumption (TCPC), the total number of rejected applications (TRA) and the total number of active wavelength (TAW) in the proposed topology as illustrated in Eq. (1). We minimize the total number of active wavelengths to conserve inactive wavelengths for future use or to improve the system's resilience.…”
Section: Model Description and Evaluation Scenariosmentioning
confidence: 99%
“…Cloud computing which offers on-demand access to computation over the Internet has become the de facto means for computing capacity consumption because of its numerous benefits. Between 2010 and 2018, the volume of compute instances in datacentres (DCs) increased by over 500% [1]. As a result, the number and size of cloud DCs has increased in the last decade.…”
Section: Introductionmentioning
confidence: 99%
“…96 vCPUs. Für diese kann man einen Stromverbrauch von einigen Hundert Watt abschätzen [39,40], was damit nicht mehr als 0,5 J je Transaktion ergibt. Dies entspricht auch der Größenordnung aus einer alternativen Abschätzung, die zumindest für ZKP-freundliche Hashfunktionen einen um den Faktor 100 erhöhten Aufwand für die Beweiserstellung im Vergleich zu einer einfachen Berechnung angibt.…”
Section: Zk-rollupsunclassified