Leveraging Fog Computing for Scalable IoT Datacenter Using Spine-Leaf Network Topology

With the internet of everything (ioe) paradigm that gathers almost every object online,huge traffic workload,bandwidth,security,and latency issues remain a concern for iot users in today's world. besides,the scalability requirements found in the current iot data processing (in the cloud) can hardly be used for applications such as assisted living systems,big data analytic solutions,and smart embedded applications. this paper proposes an extended cloud iot model that optimizes bandwidth while allowing edge devices (internet-connected objects/devices) to smartly process data without relying on a cloud network. its integration with a massively scaled spine-leaf (sl) network topology is highlighted. this is contrasted with a legacy multitier layered architecture housing network services and routing policies. the perspective offered in this paper explains how low-latency and bandwidth intensive applications can transfer data to the cloud (and then back to the edge application) without impacting qos performance. consequently,a spine-leaf fog computing network (sl-fcn) is presented for reducing latency and network congestion issues in a highly distributed and multilayer virtualized iot datacenter environment. this approach is cost-effective as it maximizes bandwidth while maintaining redundancy and resiliency against failures in mission critical applications. © 2017 k. c. okafor et al.