Cloud Elasticity for HPC Applications: Observing Energy, Performance and Cost
Vinicius Facco Rodrigues$^{1}$, Gustavo Rostirolla$^{1}$, Rodrigo da Rosa Righi$^{1}$, Cristiano André da Costa$^{1}$, Jorge Luis Victória Barbosa$^{1}$
$^{1}$Universidad de Vale do Rio dos Sinos. São Leopoldo Brazil
Schedule:Mon 19th@10:45, Room: D

Elasticity is one of the most known capabilities related to cloud computing, being largely deployed using thresholds. In this way, limits are used to drive resource mangement actions, leading to the following problem statements: How can cloud users set the threshold values to enable elasticity in their cloud applications? And what is the impact of the application’s load pattern on the elasticity? This article answers these questions for iterative high performance computing applications, showing the impact of both thresholds and load patterns on application performance and resource consumption. To accomplish this, we developed a reactive and PaaS-based elasticity model called AutoElastic and employed it over a private cloud to execute a numerical integration application. Here, we are presenting an analysis of best practices and possible optimizations regarding the elasticity and HPC pair. Considering the results, we observed that the upper threshold influences the application time more than the lower one.


	author 		= {Vinicius Facco Rodrigues and Gustavo Rostirolla and Rodrigo da Rosa Righi and Cristiano André da Costa and Jorge Luis Victória Barbosa},
	title 		= {Cloud Elasticity for HPC Applications: Observing Energy, Performance and Cost},
	booktitle 	= {2015 XLI Latin American Computing Conference (CLEI)},
	pages 		= {100--110},
	year 		= {2015},
	editor 		= {Hector Cancela and Alex Cuadros-Vargas and Ernesto Cuadros-Vargas},
	address 	= {Arequipa-Peru},
	month 		= {October},
	organization 	= {CLEI},
	publisher 	= {CLEI},
	url 		= {},
	isbn 		= {978-1-4673-9143-6},

Generated by Ernesto Cuadros-Vargas , Sociedad Peruana de Computación-Peru, Universidad Católica San Pablo, Arequipa-Perú