Dynamic Service Optimization in Distributed Cloud Networks

Distributed cloud networking enables the deployment of network services in the form of interconnected virtual network functions instantiated over general purpose hardware at multiple cloud locations distributed across the network. The service distribution problem has been previously considered as a static global optimization problem of finding the placement of virtual functions and the routing of network flows that meet a given set of demands with minimum cost. In this paper, motivated by the scale, dynamics, and heterogeneity of network services in distributed cloud networks, we address the design of distributed online solutions that provide global objective guarantees via local interactions without knowledge of the statistics of service demands. We characterize the cloud network capacity region and propose a distributed joint flow scheduling and resource allocation algorithm that stabilizes the underlying queuing system within this region, while achieving arbitrarily close to minimum average network cost, with a tradeoff in network delay. Numerical results confirm our theoretical analysis and demonstrate remarkably good convergence to the optimal cloud network configuration within a number of network settings.