An adaptive and elastic cloud based framework for service oriented computing in Internet of Things

We propose an analytical framework for service oriented computing in Internet of Thing scenarios to derive an optimal performance from an elastic extensible cloud that can serve a request from access, edge and core network. If the number of replicas at different levels (access networks, edge networks and core networks) are maintained according to the values derived from our analysis, then the drop rate and resource utilisation improves significantly without compromising the QoS constraints of the service requests. Under this design, the service consumers devices receive the services from the nearest level of replica(if any) at minimal level of cost which is the most rational form of expectation.