Is the illustration in the figure below right? IN
cloud computing: all computing, intelligence, decision and analysis happen at a central cloud.
edge computing: most of them happen at the terminal side while the central cloud do a little long-term analysis.
fog computing: most of them happen at the fog node while some long-term analysis and long-latency computing are transfered to the central cloud. The terminals are like that in cloud compuing which do not compute, so the fog node can be regared as a weak and distributed "cloud" and it is always located at a gateway.
In all of the three cases, communication networks are necessisary.
I recommend that you look at
-the Open Fog Consortium
https://www.openfogconsortium.org/
-ETSI-ISG-MEC (multiaccess edge computing)
https://www.etsi.org/technologies/multi-access-edge-computing
-Cloud Computing
http://ec.europa.eu/newsroom/document.cfm?doc_id=42968
You should look at use cases inventories and at framework architecture, to have an idea of the respective role of the three technology developments. They can be combined...
Dear Lai He , I guess we describe the main differences in our paper. You may refer to: Preprint Security Implications of Fog Computing on the Internet of Things
Dear Lai He
You are right so fog at the beginning was only for MTC, but then used widely for the most 5G network slices. Fog uses between cloud and edge devices. Edge computing is the the computing closes to end user , even in the terminal itself, or very close to it, so sometime fog computing consider as a part of edge computing with more computing resources. You can see the attached paper
There are actually a lot of contributions in the literature that deal exactly with this difference. I suggest you to have a read on the following contribution:
Conference Paper Edge and Fog Computing in Critical Infrastructures: Analysis...
To be quite frank the leading term is now Edge Computing. It has found its place in the framework standardisation of ISO-IEC-JTC1 SC42 Artificial Intelligence, where the reference architecture now buildson the report on Edge Computing in this group.
Furthermore the framework standards development at ISO-IEC-JTC1 SC41 Internet of Things also addresses Edge Computing, with sensor based systems covered. For instance Industry 4.0, Autonomous Driving are use cases of interest.
The key point is to look at nodes with each three functionalities:
-network
-compute
-data handling (store/retrieve, gather etc).
Then you revisit legacy algorithms like MapReduce, Hadoop and you split them in independent cells, which synchronise, coordinate, consolidate during execution.
It's about Distributed Processing.
An old concept of systolic algorithm becomes valuable again: like a human body, with blood flow through multiple organs (data flow) each with specific capability
Hi researcher,
We have reviewed the centralized and decentralized computing architectures in the introduction section of the paper below.
Paper: Task offloading in mobile fog computing by classification and regression tree
Article Task offloading in mobile fog computing by classification an...
- Badges
- Science method