Rancang Bangun Sistem Deteksi Dini Gangguan Link Optik Dan Automasi Validasi Redaman ONU GPON Berbasis SNMP Trap pada OLT ZTE

Maulana Nur Kholis, Ir. Budi Bayu Murti, S.T., M.T.

2026 | Tugas Akhir | D4 TEKNOLOGI JARINGAN

Keandalan jaringan Gigabit Passive Optical Network (GPON) sangat bergantung pada pengawasan infrastruktur fisik secara proaktif. Namun, sistem pemantauan konvensional yang mengandalkan metode SNMP Polling terbukti membebani kinerja server dan menyebabkan keterlambatan penanganan akibat pendekatan pemeliharaan yang masih bersifat reaktif. Oleh karena itu, Proyek Akhir ini ditujukan untuk merancang bangun sistem deteksi dini gangguan tautan optik yang terintegrasi dengan manajemen peringatan berbasis automasi pada perangkat Optical Line Terminal (OLT) ZTE.

Pengembangan arsitektur sistem dieksekusi dengan menerapkan metode SNMP Trap berbasis peristiwa (event-driven) untuk menangkap indikasi anomali secara waktu nyata (real-time). Sistem server dikembangkan menggunakan kerangka kerja FastAPI, yang dikombinasikan dengan modul automasi protokol Telnet untuk memvalidasi status operasional Optical Network Unit (ONU). Informasi peringatan kerusakan fisik, seperti Loss of Signal, Dying Gasp, dan Optical Power Drop, selanjutnya divisualisasikan melalui dashboard pemantauan Next.js dan didistribusikan sebagai notifikasi instan melalui antarmuka pemrograman aplikasi (API) Telegram.

Berdasarkan hasil pengujian, arsitektur sistem dibuktikan mampu beroperasi secara optimal dengan metrik kehilangan paket (packet loss) sebesar 0% rata-rata waktu tunda transmisi (delay) 54 milidetik. Algoritma klasifikasi anomali mencatatkan tingkat akurasi 100% menyaring peringatan palsu (false positive), dengan tingkat efisiensi waktu diagnosis otomatis mencapai 11,19 detik. Selain itu, fitur pemindaian massal (SNMP Walk) divalidasi mampu memproses 739 ONU secara stabil tanpa memicu lonjakan beban komputasi yang berarti. Melalui implementasi sistem ini, prosedur pemeliharaan jaringan dapat ditransformasikan secara efektif dari pendekatan reaktif menjadi preventif.

The reliability of a Gigabit Passive Optical Network (GPON) heavily depends on the proactive monitoring of its physical infrastructure. However, conventional monitoring systems relying on the SNMP Polling method have proven to burden server performance and cause delayed responses due to reactive maintenance approaches. Therefore, this Final Project is aimed at designing and developing an early detection system for optical link disturbances integrated with automated alarm management on ZTE Optical Line Terminal (OLT) devices.

The system architecture was developed by implementing an event-driven SNMP Trap method to capture anomaly indications in real-time. The server system was built using the FastAPI framework, combined with an automated Telnet protocol module to validate the operational status of the Optical Network Unit (ONU). Physical warning information, such as Loss of Signal, Dying Gasp, and Optical Power Drop, is subsequently visualized through a Next.js monitoring dashboard and distributed as instant notifications via the Telegram Application Programming Interface (API).

Based on the test results, the system architecture was proven to operate optimally with a 0% packet loss metric and an average transmission delay of 54 milliseconds. The anomaly classification algorithm recorded a 100?curacy rate in filtering false positives, with an automated diagnosis time efficiency reaching 11.19 seconds. Furthermore, the mass scanning feature (SNMP Walk) was validated to stably process 739 ONUs without triggering significant computational load spikes. Through the implementation of this system, network maintenance procedures can be effectively transformed from a reactive to a preventive approach.

Kata Kunci : GPON, OLT ZTE, SNMP Trap, Telnet, Sistem Deteksi Dini.