Content placement with maximum number of end-to-content paths in K-node (edge) content connected optical datacenter networks

To design survival datacenter networks, the concept of content connectivity, which is defined as the reachability of content from any point of a network after disaster failures, was recently proposed. To quantitatively measure content connectivity, we propose a novel concept of k-node (edge) content connectivity and apply it to optical datacenter networks. Furthermore, a novel disaster-resilient k-node (edge) content connected optical datacenter network (KC-ODN) is proposed in this paper. Design of the KC-ODN is divided into two sub-problems: content placement and independent end-to-content paths calculation. We consider flexible content placement (FCP), in which the content is replicated and maintained in multiple datacenters, rather than fixed content placement (FICP), in which the content is replicated and maintained in only one fixed datacenter, to realize KC-ODN. The integer linear program (ILP) model and heuristic algorithms are developed to achieve KC-ODN. Numerical results show that the method of FCP has lower wavelength consumption than the method of FICP. It is easier to realize high-level k-node (edge) content connectivity in a network with higher connectivity. Compared with the ILP method, heuristic algorithms are more scalable for large-scale optical datacenter networks.