Cluster implementation on mini Raspberry Pi computers using Round Robin Algorithm. (February 2020)

• Record Type: Journal Article
• Authors: Rohadi, E; Amalia, A; Prasetyo, A; Rahmat, M F; Setiawan, A; Siradjuddin, I
• Is Part Of: Journal of physics. Volume 1450(2020)
• Publication Date: 2020-02
• Subjects: Physics -- Congresses; 530.5
• DOI: 10.1088/1742-6596/1450/1/012068
• Languages: English
• ISSNs: 1742-6588

Abstract

Abstract: The incredible growth of the internet over the past five years makes the data traffic increase significantly. When a single server gets requests from many clients, the overload will likely occur, so requests from many clients cannot be handled optimally. This results in decreased server performance. In this work, cluster implementation on mini raspberry pi computers using the load balancing method is proposed. This system used the Round Robin Algorithm in the scheduling process in each CPU queue. The implementation of server cluster consisting of Raspberry Pi with the addition of load balancing is done to overcome the limited resources on the server. When clients in the amount of 100, 200, and 500 generated by sending 1000 requests simultaneously on each test, the response times obtained are 67.3 ms, 95.3 ms, and 182.3 ms, while throughputs are 62.50 Kbps, 76.42 Kbps, and 79.62 Kbps, respectively. The results show that the responses time and throughputs are stable with Round Robin algorithm contribution. That means the Round Robin Algorithm has availability in server clustering purposes. The system promises to be an alternative server solution that is cheaper and overcomes resource limitations on the server.