DESIGN OF CENTRALIZED ENERGY EFFICIENT VIRTUAL DESKTOP BASED COMPUTING AND COMMUNICATION MODEL (THE CASE OF ARBA MINCH UNIVERSITY)

Abstract

Adverse impact of information technology (IT) on climate change i.e. global warming are the two most challenging issues considered for research and investigation. Today green computing is the emerging domain in which such issues and challenges are under investigation and study. Universities in Ethiopia have their own data centers which consist of a number of expensive hardware, software, and other computing infrastructure resources. In modern computing environment the increase in the complexity of applications and software needs more computing power. In legacy of practices, it has been observed that increasing computing power is achieved by replacing old machines /systems/ servers with the latest high capacity systems. This incremental growth in computing and communication systems significantly increases the energy consumption by computers, servers, networks, and data centers. They are expected to continue consuming more and more energy with exponential growth. This increase in energy consumption, results in increased greenhouse gases emission and high power consumption. This also results into increase in operational cost of the ICT resources. Every desktop computer in use generates tons of carbon dioxide every year and results into a serious challenge to global warming. According to [1], Desktop computers have an unseen environmental and economic impact throughout their lifecycle. Starting from their production phase and ending with their disposal [2], approximately 15% of organizations’ energy costs and carbon footprint are from Information Technology usage out of which 39% of this value (i.e. 15%) is from desktop computers and servers. During problem investigation phase it has been observed and analyzed that institutions acquire modern desktop systems but they either fail to ensure optimum utilization of those desktop systems computing capacities or distribute those expensive and powerful systems to the users whose more focus is on non-computing intensive and communication applications. Ethiopian institutions have been still lagging behind in adoption of green strategy. The primary objective of this research was to examine the current state of ICTs in general and desktop system utilization strategies at AMU in particular for case based observation and analyze and finally develop an energy efficient centralized virtual desktop based computing &communication model. This model proposes Energy Efficient Connection Broker that can enable institutions in higher educational sector as a general and AMU as a particular case to adopt the most feasible energy efficient virtual desktop services from desktop providers. In this model energy efficiency information directory has been proposed which contains information about the CO2 emission rate and PUE. This information can be furnished by the virtual desktop service providers to declare their “green level” of services. The broker can be able to select the service provider using client’s request, type of service and preferred feature based green or energy efficiency parameter. The simulation technique was used and implemented DVFS with VM migration to support research results through experimental analysis. Simulated experiments found that the overall energy consumption of the datacenter is decreased by ~ 96 % for the given configuration of the datacenter. This technique is proved as an efficient technique to reduce the power consumption of the datacenter, as it reduces the energy consumption by approximately 96%. The use of these DVFS with VM migration experimental analysis can be assumed as significant contribution to reduce energy consumption at data centers and also can reduce the CO2 emission especially in reference to AMU ICTs. In order to save environment and its effect on humans this research can be utilized as a new knowledge to green computing and communication. The proposed can help as a decision support system for the managers, users or any educational institution like AMU for choosing the most viable energy efficient virtual desktop service providers for computing and communications.