Using Microservice Architecture as a Load Prediction Strategy for Management System of University Public Service

Abstract

The microservice architecture is widely adopted in cloud computing and the applications of software as a service (SaaS), and it can solve problems that the traditional monolithic application development cannot handle. However, new problems, for example, an unbalanced load, appear as many microservice modules coexist in one platform. These modules have complex relationships with each other, causing unavoidable performance bottlenecks. As reported in this paper, we proposed and investigated a load prediction strategy based on long short-term memory (LSTM), which is a revised neural network method, to solve these problems. We used the management system of a university’s public service as basic data in the microservice architecture and the Spring Cloud package as the experimental platform. The predicted load trend was compared with the actual load trend to prove that the proposed method can act as a reliable forecasting model. We compared the prediction results of our proposed strategy with those of other classical algorithms, including the autoregressive integrated moving average model (ARIMA), support vector regression (SVR), and LSTM, and we showed that our prediction strategy had higher efficiency than the other methods.