Intelligent Cloud Selection for Communication between OpenStack-based Multiple Clouds at TEIN using HTTP Tunneling and Observer Pattern API

This paper presents an intelligent communication mechanism and sharing resources between a Service Requesting Cloud Node (SRN) and a Service Providing Cloud Node (SPN). The proposed mechanism selects the most optimal SPN among all SPNs, deployed at TEIN (Trans Eurasia Information Network) in South Korea, Pakistan, and Malaysia. The proposed mechanism uses a Machine Learning model for selecting the most optimal SPN and HTTP tunneling with Observer Pattern API for communication. The data obtained using monitoring tools is used to train the machine learning models for selecting the most optimal node based on minimum Round-Trip Time (RTT). To this end, we utilize Extra Trees (ET), Ridge Classifier, and K-Nearest Neighbor (KNN) machine learning algorithms. The ET algorithm outperforms the other utilized algorithm with 94% accuracy. HTTP tunneling and Observer Pattern API can be used for communication between two Clouds where the Clouds support a different version of L2 tunnels, such as VXLAN or GENEVE. Access from an instance of a Cloud to the physical machine of another Cloud is provided by an HTTP tunnel using an HTTP bridge. The Observer Pattern of software engineering is implemented as an API on the target Cloud SPN to respond to the request on the HTTP bridge from the source Cloud. This way, for a specific service request, an optimal SPN is selected by ML, and communication is then established by HTTP tunneling and Observer Pattern API