R. Annisa, A. R. Makarim, M. Afif, W. E. Sulistiono, S. Ferbangkara, Ubaidah

Analisis Kinerja Layanan Cloud Computing dalam Sistem Cerdas Rekomendasi Tanaman Perkebunan

Introduction

Analisis kinerja layanan cloud computing dalam sistem cerdas rekomendasi tanaman perkebunan . Ukur kinerja cloud computing (App Engine, Compute Engine) pada sistem cerdas rekomendasi tanaman perkebunan adaptasi iklim. App Engine unggul, lebih efisien & stabil.

Abstract

Penelitian ini bertujuan untuk mengukur dan menganalisa kinerja layanan cloud computing dalam sistem cerdas rekomendasi tanaman perkebunan sebagai adaptasi perubahan iklim. Penelitian dilakukan menggunakan layanan App Engine dan Compute Engine dari Google Cloud Platform (GCP) untuk menganalisis kinerja sistem melalui pengujian beban berdasarkan parameter Response Time, Error Rate, Throughput, dan CPU Utilization. Hasil pengujian menunjukkan bahwa App Engine memiliki performa yang lebih stabil dan responsif. Pada durasi pengujian 30 hingga 300 detik, waktu respons App Engine menurun secara konsisten dari 7238 ms menjadi 252 ms. Tingkat Error Rate tetap 0% sepanjang pengujian, dengan throughput meningkat dari 11.1 hingga 73.8 req/detik. CPU utilization pada App Engine berkisar antara 31.56% hingga 59.89%, menunjukkan manajemen sumber daya yang efisien di bawah beban tinggi. Sebaliknya, Compute Engine menunjukkan fluktuasi dalam waktu respons, yang mencapai puncak 7481 ms pada pengujian 300 detik, dengan Error Rate tetap 0% tetapi throughput yang stabil di sekitar 12 req/detik. CPU utilization Compute Engine mencapai nilai maksimum mendekati 90%, menunjukkan keterbatasan dalam pengelolaan beban. Hasil ini mengindikasikan bahwa App Engine lebih direkomendasikan dalam implementasi sistem cerdas berbasis cloud mendukung keberlanjutan dan efisiensi perkebunan dalam menghadapi perubahan iklim.

Review

This study presents a timely analysis of cloud computing service performance for intelligent plantation crop recommendation systems, an increasingly critical area given the imperative for climate change adaptation. The research systematically investigates the comparative performance of Google Cloud Platform's (GCP) App Engine and Compute Engine services under varying load conditions. By measuring key performance indicators such as Response Time, Error Rate, Throughput, and CPU Utilization, the authors aim to identify the most suitable cloud environment. The findings clearly indicate a significant performance disparity between the two services, with App Engine emerging as the more stable and responsive option for this application context. The paper's strength lies in its practical relevance and empirical methodology, addressing a critical need for robust infrastructure to support intelligent agricultural systems. The systematic application of load testing to well-defined performance metrics provides concrete, quantitative data supporting the conclusions. App Engine consistently demonstrated superior performance, characterized by a substantial decrease in response times (from 7238 ms to 252 ms over 30-300 seconds), a perfect 0% error rate, and a significant increase in throughput (11.1 to 73.8 req/sec), all while maintaining efficient CPU utilization (31.56% to 59.89%). Conversely, Compute Engine exhibited greater response time fluctuations, peaking at 7481 ms, and high CPU utilization nearing 90%, signaling potential limitations in resource management under similar loads. These detailed comparisons offer valuable insights for practitioners deploying cloud-based intelligent systems. While the study provides valuable insights, there are avenues for further exploration. The scope of comparison is limited to only two specific GCP services; future work could benefit from expanding to include other relevant cloud offerings (e.g., serverless functions, container orchestration services) or different cloud providers to offer a broader comparative landscape. Furthermore, while an "intelligent system" is mentioned, the abstract does not elaborate on its specific computational demands or architectural complexity, which could influence the optimal cloud service choice beyond generic load testing. Investigating long-term performance stability, cost-effectiveness, and performance under more varied and extreme load patterns (beyond 300 seconds) would also enhance the study's practical utility and generalizability. Nonetheless, this research provides a strong foundational analysis for future optimization efforts in cloud-based intelligent agricultural systems.

Full Text

Comments

You need to be logged in to post a comment.