Manajemen adaptif komputasi awan terhadap perubahan konsumsi energi menggunakan cloudsim = Adaptive management of cloud computation to changes in energy consumption using cloudsim

Angellica, Angellica (2021) Manajemen adaptif komputasi awan terhadap perubahan konsumsi energi menggunakan cloudsim = Adaptive management of cloud computation to changes in energy consumption using cloudsim. Bachelor thesis, Universitas Pelita Harapan.

	Text (Title) Title (61).pdf Restricted to Repository staff only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (604kB)
Preview	Text (Abstract) Abstract.pdf Available under License Creative Commons Attribution Non-commercial Share Alike. Download (309kB) | Preview
Preview	Text (ToC) ToC.pdf Available under License Creative Commons Attribution Non-commercial Share Alike. Download (779kB) | Preview
Preview	Text (Chapter 1) Chapter1 (2).pdf Available under License Creative Commons Attribution Non-commercial Share Alike. Download (760kB) | Preview
	Text (Chapter 2) Chapter2.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (1MB)
	Text (Chapter 3) Chapter3.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (1MB)
	Text (Chapter 4) Chapter4.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (3MB)
	Text (Chapter 5) Chapter5.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (329kB)
Preview	Text (Bibliography) Bibliography.pdf Available under License Creative Commons Attribution Non-commercial Share Alike. Download (633kB) | Preview
	Text (Appendices) Appendices.pdf Restricted to Repository staff only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (1MB)

Abstract

Awareness of Cloud Computing services have become increasingly popular in the last decade, it requires a data center with thousands or more computer devices that consume very high energy even though they are not providing services (idle). For that we need a method or method for energy efficiency. One form that can be used is an adaptive management system for Cloud Computing. Migration of virtual machines from one server to another for energy efficiency can be done by adaptive methods. In this study, to perform the simulation, changes in virtual machine parameters, Scheduling Interval, and Host parameters are made. The analytical method used is the median absolute deviation. Where in this method a dynamic threshold calculation is carried out by detecting the use of Host over utilization. So to evaluate the analysis of the experimental results analytically with machine learning methods, namely linear regression and decision tree. The test results on the median absolute deviation method yield two recommendations. First, the comparison test results of Scheduling Interval with linear regression recommendations on power consumption. In the Scheduling Interval type # 1, there is Host type # 2 which has a high value of total Host power consumption, which is 628,99 watts/sec of the total Host power consumption of 569,82 watts/sec, meaning that this increase in total power consumption occurs at 90,59%. Second, the comparison test results between Scheduling Interval and decision tree recommendation on power consumption, has a significant ratio of 87,37% of the total Host power consumption. it means that the decision tree recommendation has succeeded in reducing the burden of energy consumption and is more efficient. As one example, the results of experiment A have an actual result of 475,00 watts/sec and a predicted decision tree of 415,00 watts/sec. In conclusion, this system proved to be adaptive and efficient, as shown in the results of the test comparisons. This can be seen in the decrease in power consumption, where previously the power consumption of non power aware Host type # 2 was 35100,00 watts/sec, and after that it was only 144,95 watts/sec. in other words, the decline that occurred was 0,41% after migration. / Mengetahui layanan komputasi awan semakin popular pada dekade belakangan ini, maka membutuhkan data center dengan ribuan atau lebih perangkat komputer yang mengkonsumsi energi dengan sangat tinggi meskipun dalam kondisi tidak memberikan layanan (idle). Untuk itu diperlukan cara atau metode untuk melakukan efisiensi energi. Salah satu bentuk yang dapat digunakan adalah sistem manajemen adaptif untuk Cloud Computing. Migrasi virtual machine dari satu server ke server lain guna efisiensi energi dapat dilakukan dengan metode adaptif. Pada penelitian ini, untuk melakukan simulasi dilakukan perubahan parameter virtual machine, Scheduling Interval, dan parameter Host. Metode analisis yang digunakan adalah median absolute deviation. Dimana pada metode ini dilakukan perhitungan threshold dynamic dengan mendeteksi penggunaan Host over utilization. Sehingga untuk melakukan evaluasi dilakukan analisa terhadap hasil percobaan secara analitik dengan metode machine learning yaitu linear regression dan decision tree. Hasil pengujian pada metode median absolute deviation menghasilkan dua rekomendasi. Pertama hasil pengujian perbandingan Scheduling Interval dengan rekomendasi linear regression pada konsumsi daya. Pada Scheduling Interval type #1 terdapat Host type #2 yang memiliki nilai total konsumsi daya Host yang cukup tinggi yaitu sebesar 628,99 watt/detik dari total konsumsi daya Host 569,82 watt/detik, artinya peningkatan total konsumsi daya Host ini terjadi sebesar 90,59%. Kedua, hasil pengujian perbandingan Scheduling Interval dengan rekomendasi decision tree pada konsumsi daya, memiliki perbandingan yang signifikan yaitu sebesar 87,37% terhadap total konsumsi daya Host. artinya rekomendasi decision tree berhasil menurunkan beban konsumsi energi dan lebih hemat. Seperti salah satu contoh hasil eksperimen A memiliki hasil aktual sebesar 475,00 watt/detik dan predicted decision tree sebesar 415,00 watt/detik. Kesimpulannya, Sistem ini terbukti adaptif dan efisien, seperti ditunjukkan pada hasil perbandingan pengujian. Hal ini terlihat pada menurunnya konsumsi daya, dimana sebelumnya konsumsi daya non power aware Host type #2 sebesar 35100,00 watt/detik, dan setelahnya hanya sebesar 144,95 watt/detik. dengan kata lain penurunan yang terjadi adalah sebesar 0.41% setelah dilakukan migrasi.

Actions (login required)

View Item