Leveraging fog and cloud computing for efficient computational offloading

摘要

Fog computing [1], [2], also known as edge computing, is an emerging Internet of Things (IoT)-centric paradigm in which traditionally cloud-based data storage, computation, and control are brought closer to the end devices. The fog layer comprises smart gateways, routers, and the end devices themselves, and can be viewed as a “descended cloud” to efficiently serve nearby clients (Figure 1). By virtue of its proximity, the fog has the potential to provide low-latency data processing for devices with stringent delay requirements, and can be coupled with the cloud's economy of scale to serve more intensive computing needs. The ability of the fog to accomodate real-time tasks makes it particularly attractive for time-sensitive applications such as wireless sensor networks and internet-connected vehicles [3]. By contrast, most cloud data centers are located near the core of the network, and therefore offer unacceptable round-trip latencies to devices at the network edge.