Design and Implementation of Multi-tenant Vehicle Monitoring Architecture Based on Microservices and Spark Streaming

Abstract

Aiming at reducing the costs for customers and making full use of software and hardware resources, this article propose a vehicle monitoring cloud architecture for multi-tenant in this paper which could provide customers with rent-and-use cloud services. This architecture is based on the springcloud microservice framework, any microservice module could be packaged into docker images and managed by kubernetes. When the system concurrency increases or decreases, the container can be dynamically expanded and recycled to rationally allocate hardware resources to solve the bottleneck of service access load. This architecture also recommends kafka message queues to decouple front-end data collection components and back-end data processing components that could achieve greater throughput. The MapReduce programming model is used to group chaotic data sent by all tenants’ vehicle terminal by tenant id, and statistical analysis is performed on the data of each tenant.