Criticality-aware service composition for cloud and network brokerage

Organizations could leverage many existing selfservice provisioning offers (compute, storage, network, security… as a service) for better flexibility and cost control. However, they are discouraged by the diversity and the complexity of composing such services with guarantees. In this paper, we propose a cloud and network broker that serves requests for end-to-end ”service packages” by composing, provisioning and stitching together in a coherent way diverse cloud and network services from different providers. Our composition algorithm takes into account multiple criteria such as consumer constraints on QoS, cost and criticality, as well as broker policies in order to determine the optimal placement. The evaluation results sustain the applicability of our approach through the efficiency and speed of the composition strategies with complex requests and large service catalogs.