Dynamic Weight Based Load Balancing for Microservice Cluster
Chang Yi, Xiuguo Zhang, Wei Cao · Proceedings of the 2nd International Conference on Computer Science and Application Engineering · 2018
In this paper1, the coarse grained load balancing algorithm, such as round based algorithm and random algorithm, which is provided by the mainstream framework of micro service cluster, is too rough, and the current proposed load balancing algorithms of fixed weights can not achieve a better load balance in the case of more micro service requests, and then the paper proposes a new method of load balancing. A new dynamic weight load balancing algorithm. This algorithm selects the data of CPU utilization rate, disk utilization rate, memory utilization rate and server connection number which can reflect the load condition of the server as the parameters of the load evaluation. This paper also selects the changes of each parameter to calculate the dynamic weight, the number of parameters and the weight of the server to evaluate the real time of the server. The load reflects the current load of the server more comprehensively and more accurately. Compared with some current load balancing algorithms, this paper dynamically compares the current load situation of each server, and finally selects the server with the smallest load to carry on the request forward, and better realizes the load balancing. According to the experimental results, we can see that the algorithm has a larger throughput and shorter response time in the case of many requests, which shows that the dynamic weight algorithm in this paper has a certain theoretical basis and practical significance.