A Study on Capabilities and Challenges of Fog Computing

Introduction

The faster maturity and stability of edge technologies (Ghahramani et al, 2017) has blossomed into a big factor in realizing scores of digitized elements / smart objects/sentient materials out of common, cheap and casual items in our midst. These empowered entities are data-generating and capturing, buffering, transmitting, etc. That is, tangible things are peppered with and prepared for the future. These are mostly resource-constrained and this phenomenon is called the Internet of Things (IoT). Further on, a wider variety of gadgets and gizmos in our working, walking and wandering locations are futuristically instrumented to be spontaneously interconnected and exceptionally intelligent in their behaviours. Thus, we hear, read and even feel connected and cognitive devices and machines in our everyday life. Once upon of a time, all our personal computers were connected via networks (LAN and WAN) and nowadays our personal and professional devices (fixed, portables, mobiles, wearables, implantables, handhelds, phablets, etc.) are increasingly interconnected (BAN, PAN, CAN, LAN. MAN, and WAN) to exhibit a kind of intelligent behavior. This extreme connectivity and service-enablement of our everyday devices go to the level of getting seamlessly integrated with off-premise, online, and on-demand cloud-based applications, services, data sources, and content. This cloud-enablement is capable of making ordinary devices into extraordinary ones. However, most of the well-known and widely used embedded devices individually do not have sufficient computation power, battery, storage and I/O bandwidth to host and manage IoT applications and services. Hence performing data analytics on individual devices is a bit difficult.

As we all know, smart sensors and actuators are being randomly deployed in any significant environments such as homes, hospitals, hotels, etc. in order to minutely monitor, precisely measure, and insightfully manage the various parameters of the environments. Further on, powerful sensors are embedded and etched on different physical, mechanical, electrical and electronics systems in our everyday environments in order to empower them to join in the mainstream computing. Thus, not only environments but also all tangible things in those environments are also smartly sensor-enabled with a tactic as well as the strategic goal of making them distinctly sensitive and responsive in their operations, offerings, and outputs. Sensors are sweetly turning out to be the inseparable eyes and ears of any important thing in near future. This systematic sensor-enablement of ordinary things not only make them extraordinary but also lay out a stimulating and sparkling foundation for generating a lot of usable and time-critical data. Typically sensors and sensors-attached assets capture or generate and transmit all kinds of data to the faraway cloud environments (public, private and hybrid) through a host of standards-compliant sensor gateway devices. Precisely speaking, clouds represent the dynamic combination of several powerful server machines, storage appliances, and network solutions and are capable of processing tremendous amounts of multi-structured data to spit out actionable insights.

However, there is another side to this remote integration and data processing. For certain requirements, the local or proximate processing of data is mandated. That is, instead of capturing sensor and device data and transmitting them to the faraway cloud environments is not going to be beneficial for time-critical applications. Thereby the concept of edge or fog computing has emerged and is evolving fast these days with the concerted efforts of academic as well as corporate people. The reasonably powerful devices such as smartphones, sensor and IoT gateways, consumer electronics, set-top boxes, smart TVs, Web-enabled refrigerators, Wi-Fi routers, etc. are classified as fog or edge devices to form edge or fog clouds to do the much-needed local processing quickly and easily to arrive and articulate any hidden knowledge. Thus, fog or edge computing is termed and tuned as the serious subject of study and research for producing people-centric and real-time applications and services.