Visible-Light Communication (VLC) has the potential to provide dense and fast connectivity at low cost. In this paper we propose SFNet, a novel VLC-enabled hybrid data center network that extends the design of wireless Data Center Networks (DCNs) into three further dimensions: (1) fully wireless at the inter-rack level; (2) no need for a centralized control mechanism on wireless links; and (3) no need for any infrastructure-level alterations to data centers. Previous proposals typically cannot realize these three rationales simultaneously. To achieve this vision, the proposed SFNet augments fat-tree by organizing all racks into a wireless small-world network via VLC links. The use of VLC links eliminates hierarchical switches and cables in the wireless network, and thus reduces hardware investment and maintenance costs. To fully exploit the benefits of the topology of SFNet, we further propose its topology design and optimization method, routing scheme, and online flow scheduling algorithm. Comprehensive experiments indicate that SFNet exhibits good topological properties and network performance.

The use of multicast transmission can efficiently reduce the consumption of network resources by jointly serving multiple destinations with a single source node. Currently, many multicast applications impose the constraint wherein multicast flows must be processed by a series of Virtual Network Functions (VNFs) before reaching their destinations. Given a multicast transmission, there are usually multiple server nodes, each of which is able to host all the required VNFs. Thus, the multicast flow should be initially steered to one or a few selected server nodes that act as pseudo sources, and the destinations will then retrieve new flow from any of these pseudo sources. In this paper, we model this kind of multicast as an uncertain multicast with multiple pseudo sources, whose routing structure is usually a forest consisting of multiple isolated trees. We then characterize and construct the Delay-guaranteed Minimum Cost Forest (D-MCF) such that each path from the source to the destination satisfies the end-to-end delay constraint. To tackle this NP-hard problem, we design two efficient methods, the Partition Algorithm (PA) and the Combination Algorithm (CA), to approximate the optimal solution. Theoretical analyses and evaluations indicate that these two methods can generate the desired routing forest for any multicast transfer. Moreover, the PA method achieves a better balance between performance and time consumption than the CA method. The evaluation results show that PA- $\mathrm{(}\mathrm{\Omega }\mathrm{+}\mathrm{20}\mathrm{)}$ can reduce total cost by $\mathrm{49.02}\mathrm{\%}$ while consuming $\mathrm{12.59}\mathrm{\%}$ more time, thus significantly outperforming the CA- $\mathrm{(}\mathrm{\Omega }\mathrm{+}\mathrm{20}\mathrm{)}$ method.