A Resource Elastic Scheduling Algorithm of Service Platform for Cloud Robotics

摘要

Cloud computing can offer the robots high performance computing resources, and the big data stored on cloud platform can enhance significantly autonomy and collaborative capacity of robots. Thus, a platform as a service level cloud platform is essential, which should provide the ability of reusing the achievements of robotics and support that multi robots can access the cloud services on demand. To ensure that the robots functional services can have a better performance of computing in high or low loads, and improve the resource utilization efficiency of service platform, there is a flexible container capacity scheduling algorithm in the resource management module. Considering the actual operation situation of different robotic services, this algorithm calculates the average load according to the CPU usage, memory usage and network usage of containers, and performs capacity elastic scheduling operation by preset upper and lower thresholds. During the period of resource flexibility expansion, users will not perceive the dynamic scheduling of functional service capacity, which ensures the stability of the platform.