Edge Intelligence : The Central Cloud is Dead – Long Live the Edge Cloud !
management through less interdependency and complexity .
An exception to this Linux predominance is found in the manufacturing industry , where various versions of the Windows operating system and Microsoft ’ s . NET framework make up the majority of implementations .
Containerization and Micro-services
A key architectural goal for edge computing should be to isolate services that can be deployed anywhere . In this case , “ anywhere ” can be on a smart device , on an IoT gateway , in a micro data center , inside a telecommunications network , in a data center , or in a public or private cloud . This isolation can be achieved through self-contained micro services or pods ( a group of services sharing resources , e . g . a database ).
With the expansion of end and edge devices , deploying and maintaining code bases will become more difficult . Containers can help mitigate this additional overhead by providing virtualization on the operating system level ..
Unlike virtual machines , containers impose little or no overhead . However , containers are less flexible than virtualization , only one operating system is used , Linux being the most popular one .
Docker has been the most popular container technology , but more recently “ rkt ” and Rocket have become widely used . The Open Container Initiative ( OCI ), a Linux Foundation project , aims to bring the competing container technologies together .
Kubernetes also introduces the concept of pods , which is a group of services deployed in containers that share certain resources , e . g . a data base , and interact via inter-process communication . Pods are an interesting concept for edge computing as they allow services to share scarce resources . Kubernetes , another open source project , is an opensource project by the Cloud Native Computing Foundation ( CNCF ) for automating deployment , scaling and management of containerized applications .
While containers are the preferred technology for deploying pods and micro-services in the cloud , they are also very applicable to edge computing and early adopters are already using containers or container-like technology on the edge . This allows the deployment of the same micro-services in the cloud , on premise and on the edge without touching the code .
It would be beneficial to edge computing to provide a uniform runtime infrastructure to deploy services both in the cloud and the edge . By deploying containers at the edge , one can optimize applications and services for bandwidth , storage , and compute power to meet the latency , scale , reliability , and ultimately cost requirements .
A commonly accepted open source platform would reduce total cost of ownership for customers and would allow solution providers to focus on
IIC Journal of Innovation - 21 -