Cindy Zefira Afiani
Fakultas Ilmu Komputer, Universitas Brawijaya

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

Implementasi Load Balancing Web Server Pada POX Controller Berbasis Penggunaan Memori Dengan Agen Psutil Pada Software Defined Network Cindy Zefira Afiani; Heru Nurwarsito
Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer Vol 5 No 1 (2021): Januari 2021
Publisher : Fakultas Ilmu Komputer (FILKOM), Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Increased requests to fulfill the needs of the user in accessing the internet occur as technology develops. The increasing number of requests is also accompanied by a very high increase in traffic which can cause overload. Web server load balancing method which is implemented in the network architecture known as software defined network can be a solution to improve the performance of the server. Several previous studies had implemented web server load balancing using algorithms like round robin, response time, and server resource based algorithms such as CPU usage based algorithm which had different performances. Resources of a server such as memory can serve as a parameter in the load balancing algorithm. The server with the smallest memory usage will service requests which are requested by the client. Memory usage based algorithm uses psutil agent to find out memory information on the server and is implemented in POX controller. The result of this study is sending requests by clients to the server with the smallest memory usage. The memory usage based algorithm is compared with the CPU usage based algorithm in terms of parameters such as throughput, response time, and memory utilization. In testing with 1800 requests as the largest connection requests, the memory usage based algorithm has a higher throughput value of 29.44 KB/s compared to the CPU usage based algorithm, which is 27.78 KB/s. Meanwhile, the memory usage based algorithm has a lower response time value of 5.68 ms compared to the CPU usage based algorithm, which is 5.88 ms. At the server with the highest memory capacity, the memory usage based algorithm has a lower memory utilization value of 12.94% compared to the CPU usage based algorithm, which is 13.20%.