Portable Packet Capture Berbasis Rule-Based Pattern Recognition untuk Klasifikasi Log Error pada Troubleshooting Jaringan di PT. NTT Indonesia Technology

Zulkarnain, M. Novan (2026) Portable Packet Capture Berbasis Rule-Based Pattern Recognition untuk Klasifikasi Log Error pada Troubleshooting Jaringan di PT. NTT Indonesia Technology. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2040221032_Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (6MB) | Request a copy

Abstract

PT NTT Indonesia Technology merupakan penyedia layanan Global Data Center yang lebih dari 56% aktivitas operasionalnya berasal dari service dan maintenance jaringan, sehingga proses troubleshooting menjadi aktivitas yang sangat penting dalam menjaga ketersediaan layanan dan menekan nilai Mean Time to Repair (MTTR). Namun, proses packet capture masih bergantung pada laptop, memerlukan waktu persiapan yang relatif lama, serta analisis hasil tangkapan masih dilakukan secara manual sehingga mengurangi efisiensi troubleshooting. Penelitian ini bertujuan merancang dan mengimplementasikan Portable Packet Capture berbasis Mini PC BKHD H31F 3L2S yang mengintegrasikan metode Port Mirroring (SPAN) dan Network TAP, memvalidasi kinerjanya berdasarkan setup time, uptime, packet loss, dan kemudahan implementasi, serta membandingkan metode Rule-Based Pattern Recognition dengan Decision Tree dalam klasifikasi hasil packet capture. Penelitian menggunakan metode Research and Development (R&D) yang meliputi analisis kebutuhan, perancangan sistem, implementasi perangkat keras dan perangkat lunak, serta pengujian komparatif. Hasil penelitian menunjukkan bahwa perangkat berhasil menurunkan setup time dari rata-rata 13,75 menit menjadi 2,65 menit, atau meningkat efisiensinya sebesar 80,7%, sehingga berpotensi mendukung penurunan MTTR pada proses troubleshooting jaringan. Pengujian juga menunjukkan perangkat mampu beroperasi mandiri dengan uptime lebih dari 2 jam, sementara metode Network TAP menghasilkan packet loss sebesar 1–5% dan Port Mirroring sebesar 2–7%, sesuai dengan target penelitian. Pada proses klasifikasi, metode Rule-Based Pattern Recognition memperoleh akurasi 99,75%, False Positive Rate kurang dari 5%, serta waktu klasifikasi rata-rata 13,14 detik, lebih cepat dibandingkan Decision Tree yang memerlukan sekitar 31 detik, dengan penggunaan memori dan konsumsi daya yang lebih rendah. Hasil tersebut menunjukkan bahwa sistem yang dikembangkan mampu meningkatkan efisiensi proses packet capture, mempercepat identifikasi gangguan jaringan, dan mendukung penurunan MTTR melalui proses klasifikasi yang cepat, akurat, dan deterministik.
=====================================================================================================================================
PT NTT Indonesia Technology is a provider of Global Data Center services, where more than 56% of its operational activities are related to network service and maintenance. Consequently, network troubleshooting plays a critical role in maintaining service availability and minimizing the Mean Time to Repair (MTTR). However, the conventional packet capture process still relies on laptops, requires a relatively long setup time, and depends on manual packet analysis, reducing the overall efficiency of troubleshooting activities. This study aims to design and implement a Portable Packet Capture system based on the BKHD H31F 3L2S Mini PC by integrating Port Mirroring (SPAN) and Network TAP methods, evaluate its performance in terms of setup time, uptime, packet loss, and implementation practicality, and compare the performance of Rule-Based Pattern Recognition and Decision Tree for packet capture classification. The study adopted a Research and Development (R&D) methodology consisting of requirements analysis, system design, hardware and software implementation, and comparative performance evaluation. The experimental results demonstrate that the proposed system reduced the average setup time from 13.75 minutes to 2.65 minutes, representing an 80.7% improvement and contributing to a potential reduction in MTTR during network troubleshooting. The system also achieved an operational uptime of more than 2 hours, while the Network TAP and Port Mirroring methods produced packet loss rates of 1–5% and 2–7%, respectively, meeting the predefined performance targets. In the classification task, the Rule-Based Pattern Recognition approach achieved 99.75% accuracy, a False Positive Rate below 5%, and an average classification time of 13.14 seconds, outperforming the Decision Tree model, which required approximately 31 seconds, while also consuming less memory and power. These results indicate that the proposed system improves the efficiency of packet capture, accelerates network fault identification, and supports MTTR reduction through a fast, accurate, and deterministic classification process.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Decision Tree, Mean Time to Repair (MTTR), Troubleshooting Jaringan, Portable Packet Capture, Rule-Based Pattern Recognition
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5103.2 Wireless communication systems. Two way wireless communication T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5105.585 TCP/IP (Computer network protocol) T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7882.P3 Pattern recognition systems
Divisions:	Faculty of Vocational > 36304-Automation Electronic Engineering
Depositing User:	M. Novan Zulkarnain
Date Deposited:	04 Aug 2026 04:29
Last Modified:	04 Aug 2026 04:29
URI:	http://repository.its.ac.id/id/eprint/142683

Actions (login required)

View Item