IMPLEMENTATION OF KUBERNETES CLUSTER USING LXD CONTAINER

Abstract

The use of cloud computing service enables clients to access data and information without having to install an application which increases the number of client requests that leads to the increase of server load and could result in overload. In clustering technology, the server workload will be distributed evenly to each server. To run clustering, a platform called Kubernetes is needed. Kubernetes is used for container management. To run a clustering with Kubernetes requires virtualization using LXD. This final project compares the performance of conventional model services and Kubernetes clusters. The test results of two servers in standby condition, the technology that uses the highest CPU and memory is the Kubernetes cluster which reached 104% on CPU usage, while memory usage is 28%. It is because of the nodes that are running on the virtual machine. In busy server conditions, conventional technology is the one that uses the highest CPU and memory due to the high, which reached 1289% on CPU usage while memory usage is 46%. Number of client requests that increase CPU and memory resources.