RUSNAIDI, SAID FAUZUL (2021) PENDETEKSIAN SERANGAN IOT BOTNET TIDAK DIKENAL DENGAN UNSUPERVISED COMPETITIVE LEARNING. S2 thesis, Universitas Mercu Buana Jakarta.
|
Text (HAL COVER)
1 Cover Said Fauzul.pdf Download (395kB) | Preview |
|
![[img]](https://repository.mercubuana.ac.id/style/images/fileicons/text.png) |
Text (BAB I)
BAB I Said Fauzul.pdf Restricted to Registered users only Download (43kB) |
|
|
Text (BAB II)
BAB II Said Fauzul.pdf Restricted to Registered users only Download (390kB) |
|
|
Text (BAB III)
BAB III Said Fauzul.pdf Restricted to Registered users only Download (436kB) |
|
|
Text (BAB IV)
BAB IV Said Fauzul.pdf Restricted to Registered users only Download (4MB) |
|
|
Text (BAB V)
BAB V Said Fauzul.pdf Restricted to Registered users only Download (32kB) |
|
|
Text (DAFTAR PUSTAKA)
Daftar Pustaka Said Fauzul.pdf Restricted to Registered users only Download (146kB) |
|
|
Text (LAMPIRAN)
Lampiran Said Fauzul.pdf Restricted to Registered users only Download (228kB) |
Abstract
The use of IoT devices as a Robot Network (IoT Botnet) is a serious threat to cybersecurity. This is inseparable from the growth of IoT devices equipped with advanced communication technology and computing systems, but not accompanied by a strong security system in place. To anticipate the rapid growth of IoT devices and/or types of Botnets in the future, research is needed to detect IoT Botnets on new IoT devices and/or new types of Botnets (Unknown Attack) which have never been involved in the training process. In this study, We proposed to detect unknown IoT Botnets using the Unspervised Learning approach which is applied to Competitive Learning based method (hereinafter referred to as Unsupervised Competitive Learning). The test was carried out using Learning Vector Quantization (LVQ) method and the N-BaIoT dataset: Data for network based detection of IoT botnet attacks, which provides real traffic data from 9 commercial IoT devices that have been infected with Bashlite/Gafgyt and Mirai Botnet. Testing was carried out in 257 batches divided into 5 test scenarios, namely detection of IoT Botnet using combined data from all IoT devices and all kind of IoT Botnets as baseline scenario (Scenario-0) with the highest accuracy of 99.54%, detection of known IoT Botnets on known IoT devices (Scenario-1) with an average total accuracy of 85.22%, detection of known IoT Botnets on unknown IoT devices (Scenario-2) with an average total accuracy of 83.06%, detection of unknown IoT Botnet on known IoT devices (Scenario-3) with an average total accuracy of 79.98% and detection of unknown IoT Botnets on unknown IoT devices (Scenario-4) with an average total accuracy of 75.07%. Keywords: IoT Botnet, Unknown Attack, Unsupervised Learning, Competitive Learning, Learning Vector Quantization. Pemanfaatan perangkat IoT sebagai Robot Network (IoT Botnet) merupakan ancaman serius pada cybersecurity. Hal tersebut tidak terlepas dari pertumbuhan perangkat IoT yang dilengkapi dengan teknologi komunikasi dan sistem komputasi yang canggih, namun tidak diiringi dengan kuatnya system keamanan yang diterapkan. Untuk mengantisipasi cepatnya pertumbuhan perangkat IoT dan/atau jenis Botnet di waktu yang akan datang, dibutuhkan penelitian untuk pendeteksian IoT Botnet pada perangkat IoT dan/atau jenis Botnet baru yang tidak dikenali sebelumnya (Unknown Attack). Dalam penelitian ini diusulkan pendeteksian IoT Botnet tidak dikenal menggunakan pendekatan Unspervised Learning yang diterapkan pada metode pembelajaran berbasis Competitive Learning (pada penelitian ini disebut dengan Unsupervised Competitive Learning). Pengujian dilakukan dengan metode Learning Vector Quantization (LVQ) dan memanfaatkan dataset N-BaIoT: Data for network based detection of IoT botnet attack, yang menyediakan data real traffic dari 9 perangkat IoT komersial yang telah diinfeksi Bashlite/Gafgyt dan Mirai Botnet. Pengujian dilakukan dalam 257 batch yang terbagi pada 5 skenario pengujian, yaitu pendeteksian IoT Botnet yang menggunakan data gabungan dari keseluruhan perangkat IoT dan IoT Botnet sebagai baseline scenario (Skenario-0) dengan nilai akurasi tertinggi didapatkan sebesar 99,54%, pendeteksian IoT Botnet sudah dikenal pada perangkat IoT sudah dikenal (Skenario-1) dengan rata-rata total akurasi didapatkan sebesar 85,22%, pendeteksian IoT Botnet sudah dikenal pada perangkat IoT belum dikenal (Skenario-2) dengan rata-rata total akurasi didapatkan sebesar 83,06%, pendeteksian IoT Botnet belum dikenal pada perangkat IoT sudah dikenal (Skenario-3) dengan rata-rata total akurasi didapatkan sebesar 79,98% dan pendeteksian IoT Botnet belum dikenal pada perangkat IoT belum dikenal (Skenario-4) dengan rata-rata total akurasi didapatkan sebesar 75,07%. Kata Kunci: IoT Botnet, Unknown Attack, Unsupervised Learning, Competitive Learning, Learning Vector Quantization
| Item Type: | Thesis (S2) |
|---|---|
| NIM/NIDN Creators: | 55419110011 |
| Uncontrolled Keywords: | IoT Botnet, Unknown Attack, Unsupervised Learning, Competitive Learning, Learning Vector Quantization |
| Subjects: | 600 Technology/Teknologi > 620 Engineering and Applied Operations/Ilmu Teknik dan operasi Terapan 600 Technology/Teknologi > 620 Engineering and Applied Operations/Ilmu Teknik dan operasi Terapan > 621 Applied Physics/Fisika terapan |
| Divisions: | Pascasarjana > Magister Teknik Elektro |
| Depositing User: | Dede Muksin Lubis |
| Date Deposited: | 23 Oct 2023 08:01 |
| Last Modified: | 23 Oct 2023 08:01 |
| URI: | http://repository.mercubuana.ac.id/id/eprint/83097 |
Actions (login required)
 |
View Item |
