Service Dependency Based Dynamic Load Balancing Algorithm for Container Clusters

Recently e-business have grown explosively with the widespread of Internet of Things (IoT). To acheive better performance and higher resource utilization, many services providers design their services following the principle of micro service and deploy them on container clusters like kubernetes. However, the existing approaches fail to address service dependency problem. The service dependency is a challenge introduced when micro service is applied for e-business, the performance of a service is not only decided by the processing ability of its backend servers, but also by performance of other serivces it relies on to provide its functionality. To address this, this paper proposes a dynamic load-balancing solution for kubernetes which can combine service dependency to give a more accurate description for request resource consumption and backend server ability, and finally performs a better load balancing. Experimental results are given to validate the performance of proposed solution.