Electronic Theses and Dissertation

Protokol Message Queuing Telemetry Transport (MQTT) merupakan protokol komunikasi yang banyak digunakan dalam sistem Internet of Things (IoT) karena sifatnya yang ringan dan efisien. Namun, kinerjanya dapat dipengaruhi oleh beberapa faktor, salah satunya panjang topik (topic length) yang digunakan dalam proses publikasi dan langganan pesan. Penelitian ini mencakup dua skenario utama, yaitu growth test dan metrics test. Growth test mengevaluasi kapasitas broker terhadap peningkatan jumlah klien, sedangkan metrics test menganalisis kinerja komunikasi berdasarkan empat parameter utama yaitu Message Transmission, throughput, latency, dan packet loss. Variasi panjang topik yang diuji meliputi 5, 10, 20, 50, dan 100 karakter, dengan jumlah klien pada pengujian metrics berturut-turut 100, 200, 300, 400, dan 500. Hasil growth test menunjukkan bahwa konsumsi RAM mencapai sekitar 95% dengan penggunaan CPU hanya 5-13%, menandakan bahwa beban utama broker berada pada sisi memori, bukan komputasi. Sistem mencapai batas maksimum pada kisaran 90-112 klien dengan total 6.329 paket terkirim, yang menunjukkan bahwa kapasitas broker lebih dipengaruhi oleh ketersediaan memori daripada kecepatan prosesor. Pada metrics test, panjang topik memberikan pengaruh terhadap Message Transmission sebesar 38,58%, throughput 11,98%, latency 12,03%, dan packet loss 19,47%. Nilai rata-rata Message Transmission sebesar 54.242,40 bytes/s, throughput 189,36 msg/s, latency 0,0505 s, dan packet loss 0,0176% menunjukkan bahwa MQTT tetap stabil dan andal di berbagai panjang topik. Secara keseluruhan, penelitian ini menyimpulkan bahwa panjang topik memiliki pengaruh terbatas terhadap performa komunikasi MQTT. Faktor yang lebih dominan terhadap kestabilan sistem adalah jumlah klien dan kapasitas memori broker. Dengan demikian, peningkatan kapasitas RAM dan manajemen jumlah klien menjadi aspek utama dalam mengoptimalkan performa MQTT pada implementasi IoT berskala besar. Kata Kunci: MQTT, Internet of Things, Topik MQTT, Message Transmission, Throughput, Latency, Packet loss, Broker, CPU, RAM.

The Message Queuing Telemetry Transport (MQTT) protocol is a widely used communication protocol in Internet of Things (IoT) systems due to its lightweight and efficient nature. However, its performance can be affected by several factors, one of which is the topic length used in the message publishing and subscription process. This study covers two main scenarios the growth test and the metrics test. The growth test evaluates the broker’s capacity concerning the increasing number of clients, while the metrics test analyzes communication performance based on four main parameters, namely Message Transmission, throughput, latency, and packet loss. The tested topic length variations include 5, 10, 20, 50, and 100 characters, with the number of clients in the metrics test being 100, 200, 300, 400, and 500, respectively. The growth test results show that RAM consumption reached around 95%, with CPU usage only between 5-13%, indicating that the broker’s primary load lies on memory rather than computation. The system reached its maximum limit at around 90-112 clients with a total of 6,329 packets sent, demonstrating that the broker’s capacity is more influenced by memory availability than processor speed. In the metrics test, topic length affected Message Transmission by 38.58%, throughput by 11.98%, latency by 12.03%, and packet loss by 19.47%. The average values of Message Transmission (54,242.40 bytes/s), throughput (189.36 msg/s), latency (0.0505 s), and packet loss (0.0176%) indicate that MQTT remains stable and reliable across various topic lengths. Overall, this study concludes that topic length has a limited impact on MQTT communication performance. The more dominant factors affecting system stability are the number of clients and the broker’s memory capacity. Therefore, increasing RAM capacity and managing the number of clients are key aspects in optimizing MQTT performance for large-scale IoT implementations.