Figure 2

Cloud system modelling. Design of the proposed cloud architecture. User requests from multiple devices go through a HTTP interface to the cloud system. A First-Come-First-Serve scheduler distributes all these requests to the Front-end nodes, which forward these to the Back-end nodes. The Back-end nodes process the requests and compute the expected user result, accessing the system database if needed. In the Back-end, there are also control nodes that monitor the state of the system, and are able to create or destroy virtual machines according to that state.