AI on the Edge with Security

Abstract

We, Information Technologists, are adept at coining picturesque names for new technological innovations.In the early part of this century, we saw the emergence of huge data centers established by vendors that were accessible to customers via the Internet from any place at any time.They provided on-demand and payment services such as computing cycles, data storage, computing platforms, and software systems.They were multi-tenanted, elastic, and virtualized.Instead of calling them a "Computing Utility on the Internet," the term Cloud Computing was coined befuddling non-professionals who wondered how Clouds could compute!Cloud Computing was inevitably followed by Fog Computing, Mist Computing, and Dew Computing-more picturesque names for cloud extensions.These computing environments are visualized as layers: CloudFogMist and Dew.In between Dew and Mist, a more ordinarily named environment-Edge Computing-has emerged.It is a computing layer at the edge of the Internet that can access the cloud while processing locally critical data of systems they govern.The most important properties of Edge Computing are low latency and high security, both properties vital to process data acquired from a plethora of IoT devices.The data acquired from these devices are securely processed primarily by Edge computers.Novel ideas from the emerging Artificial Intelligence paradigms are used to solve problems arising from the growing applications of computers in health care systems, manufacturing systems employing armies of robots, logistics in moving goods, and plenty of novel systems.Security and Intelligence are the two vital requirements in emerging computer-assisted systems and Edge Computing provides them.