Avesha Smart Application Load Balancer Partner Validated

Load Balancing (LB) has traditionally been based on Weighted Round-Robin (WRR) or some variants of it. This has worked quite well with suitable auto-scaling and over-provisioning of resources. However, this is posing a problem in the present situation with microservices workloads deployed on hundreds and even thousands of cloud compute resource instances resulting in a large amount of wastage. Hand tuning of weights is not practical especially with dynamic workloads that can vary in minutes. Avesha solves this issue using automation based on Reinforcement Learning (RL) – a machine learning technology that learns to autonomously optimize the weights to achieve a desired performance while minimizing waste.