A cloud-edge collaboration messaging system for edge computing

Abstract:
Message Queue is an important tool for data exchange. In the edge computing domain, there are lots of data exchange requirements between Cloud and Edge, which brings some new challenges to Message Queues, such as the limit of memory resource, the lack of security protection in the edge, the unstable network between Cloud and Edge, and so on. In this paper we propose a cloud-edge collaboration messaging system for edge computing. It is a cloud native messaging system, of which all components are implemented and deployed by containers, and can be operated by compatible Kubernetes APIs. It includes a messaging cluster on the cloud side to provide the high reliability and scalability of Message Queue by the power of central cloud, and a lite messaging service on each edge side to provide a buffer for messages from Edge to Cloud. Moreover, it uses authentication and segregation to enhance the security of the communication between Cloud and Edge. This messaging system not only provides high security, reliability and scalability for the communication between Cloud and Edge, but also consumes very low memory and CPU resources. According to the performance comparison testing, it has an outstanding throughput and latency value, especially for the communication between Cloud and Edge of the edge computing domain with a large scale of connections.