Energiasäästliku uduandmetöötluse simuleerimine
Date
2019
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Nõudlus arvuti ressursside järele üha suureneb ning seega on vajadus vähendada energiakulu, et tagada arvutisüsteemide jätkusuutlikus. Praegused pilve- ja uduandmetöötlus arhitektuuride edasiarendamiseks on vaja ajajaotus- ja asetusalgoritme, mis arvestavad energiakuluga. Selles töös kirjeldatakse energiasäästlikkust pilve- ja uduandmetöötluses. Töös luuakse ajajaotus- ja asetusalgoritmid, mis maksimeerivad vabade seadmete arvu ning vähendavad seeläbi süsteemi energiakulu. Algoritme katsetatakse erinevates simulatsioonides. Simulatsioonide tulemusi analüüsitakse ja võrreldakse ning tehakse järeldused algoritmide kasulikkusest. Töö sisaldab ka lühikest ülevaadet sarnastest uurimustest.
With increasing demand on computing resources, there is a need to reduce energy consumption in order to keep computer systems sustainable. Current cloud and fog computing architectures need to be improved by designing energy efficient scheduling and placement algorithms. This thesis describes power efficiency in fog computing and cloud computing. It shows a way to minimize power usage by designing scheduling and placement algorithms that maximize the number of idle hosts. Algorithms are designed to archive that goal in cloud and fog systems. The algorithms are tested in different simulation scenarios. The results are compared and analysed. The thesis also contains a brief overview of similar research that has been done on this topic.
With increasing demand on computing resources, there is a need to reduce energy consumption in order to keep computer systems sustainable. Current cloud and fog computing architectures need to be improved by designing energy efficient scheduling and placement algorithms. This thesis describes power efficiency in fog computing and cloud computing. It shows a way to minimize power usage by designing scheduling and placement algorithms that maximize the number of idle hosts. Algorithms are designed to archive that goal in cloud and fog systems. The algorithms are tested in different simulation scenarios. The results are compared and analysed. The thesis also contains a brief overview of similar research that has been done on this topic.