Q: What is edge computing and where is the edge?
A: Data collected from early Internet of Things (IoT) applications was processed in a cloud infrastructure with decisions pushed back from the cloud to things. While this architectural paradigm is suitable for a subset of consumer IoT, it quickly shows its limitations in the context of industrial IoT.
Edge computing is a viable alternative to these cloud-centric architectures which enables data to be processed closer to where it is produced and can be as far as the device itself. Therefore, the edge is not a definitive line and should be drawn wherever it is needed by different applications. ADLINK has been one of the driving forces of this new technology and a key contributor to the Eclipse EDGE Native working group and fog05 project.
Q: Is there a difference between edge and fog computing?
A: The increasing awareness of the challenges posed by cloud-centric architectures (such as connectivity, latency, throughput, cost and security), motivated the development of edge and fog computing as a decentralized/distributed and fluid alternative.
This question is often asked and does cause some confusion but essentially the two terms have converged to become the same thing. The names came from different sectors with Fog and Mist inspired by the original term cloud. At the other end, the telecommunications industry first used the term edge computing in the context of
multi-access edge computing (MEC) which became generalised and adopted by other sectors. Today edge and fog are synonymous.
Q: What is fog05 and its use cases?
A:
Eclipse fog05 is an open source project aimed at providing a platform designed ground-up for addressing the industry’s most pressing needs for edge and fog computing. The platform supports seamless, decentralized management, provisioning and monitoring of infrastructure distributed across the Internet, spanning almost any system from data centre servers to micro-controllers.
The first fog05 release in April 2020 delivered a stepping-stone for accelerating the adoption of innovative edge and fog computing architectures. It enables users to leverage the power of decentralization in order to exploit local computation, maintain data where it makes the most sense – and push the edge further down than ever before. It also reduces the overall energy consumption of the system, thus contributing to a better energy footprint.
It’s designed for anyone needing to build and manage a system which embraces edge architecture. For example: ETSI MEC and ETSI NFV convergence; fog-assisted robotics; vehicle-to-everything (V2X); augmented reality/virtual reality; and industrial IoT applications in general.
Q: Why is fog05 different?
A: The first release of Eclipse fog05 is an important landmark in terms of enabling the ecosystem and underlining that full decentralization is the way forward.
It is the only platform designed from scratch which focuses on fog and edge computing. Largely based on ADLINK technology, it addresses highly heterogeneous systems even those with extremely resource-constrained nodes; offers a common API and information model for control; and unifies the virtualization of compute, networking and storage infrastructure at the deepest edge.
By eliminating dependency on special nodes and servers, users can maintain full control of their system and data despite failures and network partitions. In addition, as fog05 is an Eclipse project, users can be confident that it is safe to use, does not violate any patents and offers content permanence. It can only be accessed via the official eclipse licensing scheme and has been recognized by ETSI as being compliant with MEC/NFV.
Q: How do you see edge/fog computing developing?
A: The first step was to make available the foundations of an infrastructure specifically designed for edge computing. Thanks to fog05 those foundations are now in place and over the next 12-24 months we can expect an acceleration in the number of edge and fog applications. At ADLINK, we will continue to work hard building edge services and advanced scheduling, and allocation algorithms.
At the same time, the differences between the edge and cloud-centric approaches will become increasingly evident and hence encourage faster adoption of decentralized architecture. Those who produce the data should own it and decide how it can be used for analytics and decentralization and by whom. By making ownership extremely clear, we hope to make it easier for data monetisation at the point of creation, rather than a prerequisite of the centralised collecting parties.
Read the press release to learn more about fog05 or visit us on
GitHub.
Originally published in German here:
https://webkiosk.epaper-kiosk.beam-verlag.de/8-2020/63601233/fullscreen/18