A Survey on Energy Efficiency in Information Systems

Abstract

Received (Day Month Year) Revised (Day Month Year)Concerns about energy and sustainability are growing everyday involving a wide range of fields. Even Information Systems are being influenced by the issue of reducing pollution and energy consumption and new fields are rising dealing with this topic. One of these fields is Green IT, which deals with energy efficiency with a focus on IT. Researchers have faced this problem according to several points of view. The purpose of this paper is to understand the trends and the future development of Green IT by analyzing the state of the art and classifying existing approaches to understand which are the components that have an impact on energy efficiency in Information Systems and how this impact can be reduced. At first, we explore some guidelines that can help to understand the efficiency level of an organization and of an Information System. Then we discuss measurement and estimation of energy efficiency and identify which are the components that mainly contribute to energy waste and how it is possible to improve energy efficiency, both at the hardware and at the software level.

A Taxonomy and Future Directions for Sustainable Cloud Computing

“…Orgerie et al [57] presented a survey on improving the energy-efficiency of large scale distributed systems. A broad review of energy efficiency in information systems at the hardware and at the software level is offered by Vitali and Pernici [58]. Shuja et al [71] reviewed the techniques and architectures for designing energyefficient datacenters and presented an overview of server cooling and storage component.…”

Characterizing Energy per Job in Cloud Applications

“…The surveys of [4][5][6] describe different approaches focusing on multiple aspects of energy consumption, which are not limited to the IT infrastructure, but also take into account the physical infrastructure and cooling.…”

“…In general, this can lead to a less performant system compared to round robin, but this modification allows the models presented in Figures 3 to 5 to become separable networks and be analyzed in linear time O(n) using the mean value analysis technique [30]. Figure 21 shows the results for different values of α ∈ {0, 25%, 50%, 75%, 100%}, with the number of VMs in the range [4,24]. We have avoided the cases with n < 4, because we wanted n slow and n f ast to be different integers for all of the considered values of α.…”

“…We will present results where newer machines in the round robin cycle are selected first (we also have tested the case in which priority is given to slower machines and obtained basically the same behavior). Figure 22 shows the results for α ∈ {0, 25%, 50%, 75%, 100%} and the number of VMs in the range [4,24]. As a general trend, we see that there is always a noticeable performance increase, both in terms of response time and energy per job reduction, whenever the number of VMs allows terminating the N jobs in one less cycle: with N = 50 jobs, this occurs with the number of VMs n ∈ {1-10, 13, 17, 25, 50}.…”

Unveiling Correlations via Mining Human-Thing Interactions in the Web of Things

“…While such a ubiquitous WoT environment offers the capability of integrating information from both the physical world and the virtual one leading to tremendous business and social opportunities (e.g., efficient supply chains, independent living of elderly persons, and improved environmental monitoring), it also presents significant challenges [Ferscha 2012;Baresi et al 2015;Vitali and Pernici 2014;Sheng et al 2008]. With many things connected and interacted over the Web, there is an urgent need to efficiently index, organize, and manage these things for object search, recommendation, and mash-up, and effectively reveal interesting patterns from things.…”