Pemanfaatan Docker Swarm dan OpenFaaS untuk Implementasi Model Komputasi Serverless pada Sistem Cloud Scalable

Maulana, Rifqi Akhmad (2025) Pemanfaatan Docker Swarm dan OpenFaaS untuk Implementasi Model Komputasi Serverless pada Sistem Cloud Scalable. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5027211035-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Penelitian ini membandingkan performa arsitektur serverful dan serverless dalam menghadapi beban tinggi, dengan variasi jumlah replika dan jenis pemrosesan request (synchronous dan asynchronous). Evaluasi mencakup tingkat error, waktu respons, serta efisiensi penggunaan CPU dan memori pada gateway/server. Hasil menunjukkan bahwa pada skenario satu instance, serverless asynchronous mencatat nol error namun dengan waktu pemrosesan lebih lambat. Serverless synchronous menunjukkan error rate dan waktu respons lebih tinggi masing-masing sebesar 16.66% dan 15.86% dibanding serverful. Sistem dengan lebih banyak replika, khususnya serverless dengan 8 dan 10 replika, memberikan performa lebih stabil. Dibanding sistem ini, serverful satu worker lebih cepat 13.83%, dan serverless satu replika lebih lambat 41.84%, namun keduanya mengalami error lebih dari 2300% relatif terhadap serverless 10 dan 8 replika. Pada pengujian concurrency, penambahan instance pada serverful menurunkan error hingga 75.09%. Skor Apdex tertinggi dicapai oleh serverful lima instance dan serverless dengan auto-scaling. Namun, waktu cold start pada serverless rata-rata 300 kali lebih lambat dibanding warm start. Temuan ini diharapkan dapat menjadi rujukan dalam memilih arsitektur cloud yang optimal sesuai dengan kebutuhan aplikasi.

=======================================================================================================================================

This study discusses the performance comparison of serverful and serverless architectures under various high-load conditions, with variations in the number of replicas and request processing types (synchronous and asynchronous). The evaluation covers error rate, response time, and CPU/memory usage efficiency on the gateway/server. Results show that in a single-instance scenario, serverless asynchronous recorded zero errors but had significantly slower processing time. Serverless synchronous exhibited 16.66% higher error rate and 15.86% slower response time compared to serverful. Systems with more replicas, particularly serverless with 10 and 8 replicas, demonstrated more stable performance. Compared to this setup, serverful with a single worker was 13.83% faster, and serverless with one replica was 41.84% slower, with both experiencing over 2300% more errors relative to serverless with 10 and 8 replicas. In concurrency testing, increasing serverful instances reduced errors by up to 75.09%. The highest Apdex scores were achieved by serverful with five instances and serverless with auto-scaling. However, serverless cold start time was on average 300 times slower than warm start. These findings are expected to serve as a reference in selecting the optimal cloud architecture based on application needs.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Serverful, Serverless, Asynchronous, Apdex, Cold start, Warm start, replika
Subjects:	T Technology > T Technology (General) T Technology > T Technology (General) > T58.64 Information resources management T Technology > T Technology (General) > T58.8 Productivity. Efficiency
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information Technology > 59201-(S1) Undergraduate Thesis
Depositing User:	Rifqi Akhmad Maulana
Date Deposited:	01 Aug 2025 02:53
Last Modified:	01 Aug 2025 02:53
URI:	http://repository.its.ac.id/id/eprint/125477

Actions (login required)

View Item