Analisa Prediksi Migrasi Pada Virtual Machine Server Menggunakan Metode Decision Tree

AZZERY, YASEP (2021) Analisa Prediksi Migrasi Pada Virtual Machine Server Menggunakan Metode Decision Tree. S2 thesis, Universitas Mercu Buana Jakarta.

Preview	Text (HAL COVER) 01 Cover - Yasep Azzery.pdf Download (540kB) | Preview
	Text (BAB I) 02 Bab 1 - Yasep Azzery.pdf Restricted to Registered users only Download (200kB)
	Text (BAB II) 03 Bab 2 - Yasep Azzery.pdf Restricted to Registered users only Download (453kB)
	Text (BAB III) 04 Bab 3 - Yasep Azzery.pdf Restricted to Registered users only Download (264kB)
	Text (BAB IV) 05 Bab 4 - Yasep Azzery.pdf Restricted to Registered users only Download (458kB)
	Text (BAB V) 06 Bab 5 - Yasep Azzery.pdf Restricted to Registered users only Download (174kB)
	Text (DAFTAR PUSTAKA) 08 Daftar Pustaka - Yasep Azzery.pdf Restricted to Registered users only Download (344kB)
	Text (LAMPIRAN) LAMPIRAN Yasep Azzery.pdf Restricted to Registered users only Download (426kB)

Abstract

Virtual technology Server machines work by emulsifying physical devices that are presented into virtual devices, several components including CPU, Memory, and Network. Performance of each VM server component in each host when carrying out its functions must always be balanced so that there is no excess that causes downtime. Several VM migration techniques are applied to balance the server performance load. In this study use Decision Tree classification method with the C.45 algorithm to analyze data and predict VM migration by classifying performance of each component. VM data classification is divided into CPU, memory, and network performance data which are categorized into low, medium, and high performance. The classification is determined based on the range of performance values from the VM aspect. Prediction of migration and non-migration of VMs is done by analytical calculations to determine the threshold value for VM migration and migration categories. Testing with the Decsion Tree method using the C.45 algorithm produces a gain value for the CPU of 0.52249, a memory gain value of 0.00890, and a network gain of 0.37038. Based on the gain value obtained, CPU atribut is the root node of the decision tree that is formed. Results of the comparison test of migration and non-migration analysis carried out by analytical calculations, compared with predictions of the Decision Tree method resulted in an accuracy value of 97.48%. Teknologi Virtual Mesin server bekerja dengan cara mengemulsi perangkat fisik yang di presentasikan kedalam perangkat virtual, beberapa komponen diantaranya CPU, Memory, dan Network. Performa setiap komponen VM server dalam tiap host ketika menjalankan fungsinya harus selalu seimbang agar tidak kelebihan yang menyebabkan downtime. Beberapa teknik migrasi VM diterapkan untuk menyeimbangkan beban kinerja server. Metode yang digunakan dalam penelitian ini yaitu metode klasifikasi Decision Tree dengan algoritma C.45 untuk menganalisis data dan melakukan prediksi migrasi VM dengan mengklasifikasikan kinerja masing-masing komponen. Klasifikasi data VM terbagi atas data performa CPU, memory, dan network yang dikategorikan kedalam klasifikasi rendah, sedang, dan tinggi. Klasifikasi ditentukan berdasarkan range nilai performa dari aspek VM tersebut. Prediksi migrasi dan non migrasi VM dilakukan dengan perhitungan analitis untuk menentukan nilai ambang batas kategori migrasi dan migrasi VM. Pengujian dengan metode Decsion Tree menggunakan C.45 algorithm menghasilkan nilai gain untuk CPU sebesar 0,52249, nilai gain memory sebesar 0,00890, dan Network sebesar 0,37038. Berdasarkan nilai gain yang diperoleh, CPU merupakan root node dari pohon keputusan yang terbentuk. Hasil pengujian perbandingan analisa migrasi dan non migrasi yang dilakukan dengan perhitungan analitis, dibandingkan dengan prediksi metode Decision Tree menghasilkan nilai akurasi sebesar 97,48%.

Actions (login required)

View Item