Weight based Load Balancing in Kubernetes using AWS

Weight based Load Balancing in Kubernetes using AWS

연구 분야: Software Development

학회: 2023 International Conference on Intelligent Data Communication Technologies and Internet of Things (IDCIoT)

Kubernetes is a container orchestration system to automate software deployment,scaling and management of containers. Kubernetes offers Load Balancer service to distribute external traffic to the pods.Load balancing is the process of distributing network traffic across multiple servers and prevent a single server from being overloaded. This article has discussed about the existing load balancing algorithms and different performance met that affects the load balancing. Canary releases helps to deploy applications or services to a small percentage of users which minimizes the scope, impact, and risk associated with introducing new software assets into production.To implement a weight-based load balancer in Kubernetes using AWS, an healthy Kubernetes cluster is deployed in the AWS. An automated canary deployment approach is utilized, to distribute traffic based on weights assigned to the services. Linkerd and Flagger traffic split functionality helps to dynamically switch the traffic across services. Deployed sample applications, primary version and canary version in the cluster and configured weights of the application. The load balancer shifts the traffic from 10% ,100% with the step size of 10% to Canary version. All the target environment can be updated with small increments such as 10%, 25%, 75%, and 100% traffic load increments.