By Qi Wang
Picture this: all of your devices are interconnected – your smartphone, tablet, laptop, car, television set, RoboVac, and even your air conditioner. They communicate with each other, command your needs to each other, and are at your whim and fancy with a simple voice demand. This is the idea behind the Internet of Things (IoT), where all devices are interconnected. The connectivity requires in one tenth of a millisecond megabyte of information to be transmitted, processed, analyzed and stored. This is carried out by a powerful big data processing system called “cloud computing”. In this system, a central server collects data from end devices, analyzes these data, draws conclusions, and sends feedback to where they are needed.
As amazing as it sounds, issues do exist. For instance, the capacity the system takes, the latency it might cause, and one point of failure that will potentially paralyze or collapse the entire system. Hence, the highly centralized system needs to be decentralized. This is why fog and edge computing exist.
The computing spectrum
Think of it as a spectrum. Cloud computing is highly centralized, whereas fog is less centralized and more localized, and edge is highly localized. Fog computing uses multiple data processing centers instead of one, pushing data processing closer to end devices. And edge computing enables end devices to independently process information and decide for themselves what information to store in its own system and what information to send up the chain.
Fog vs. edge computing
Imagine a metropolis’s road network, and all vehicles are autonomous and interconnected. They communicate traffic information with the rest of the network from their vicinity. All the information could be sent to and processed in one network-wide centralized server named the “cloud”, and then routing commands sent back from the cloud to the vehicles. But latency and potential failures in this highly centralized system would probably make a passenger curse.
Alternatively, data can be transmitted and processed via multiple servers also known as “fogs” spread across the network. Burden on each server and latency would be dramatically reduced. Servers would transmit up the chain only the information they deem significant to the whole network. Even when one server fails, others would still function and it would not bring down the whole network. Furthermore, in an edge computing network, individual vehicles are able to process traffic information within the vicinity and independently make routing decisions. Only information beneficial to a wider network would be transmitted up the chain, like information on a car crash of severe magnitude.
Gaining an ‘edge’ in China
In China, edge computing is gaining more ground. The Edge Computing Consortium (ECC) was established on November 30, 2016 in Beijing to fuel developments in this area. Among the big name players are HUAWEI, Intel, ARM, and government agencies. As young as it is, the ECC has published a white paper outlining its framework and three development goals – connectivity, artificial intelligence and automation. Detailed plans with regard to experimentation and application have yet to come. Government, state-owned enterprises and state controlled industries are the biggest tech consumers in China. Most technological advancements are funded by the government. So with the government’s participation and funding, developments in edge computing will likely have a strong momentum in China.
Interestingly, Intel and ARM are also involved in the Open Fog Consortium (OpenFog) along with other heavyweights like Cisco, Microsoft and Dell. As the name indicates, OpenFog represents the latest developments in fog computing. Compared to OpenFog, the ECC is still the new kid on the block. OpenFog’s website is flooded with information on the newest developments. It has formed international alliances, published multiple white papers, and made real-world applications such as high-scale package drone delivery. Moreover, lots of events are scheduled to take place in the year of 2017. So, the ECC has a lot of catching up to do.
China’s ‘edge’ vs. the ‘fog’ in the West
Several reasons might explain why China has gone ahead with edge computing when the West*, especially Silicon Valley, is focusing on fog computing. Since fog computing is regionally centralized, it offers a better big picture view and therefore better scalability in comparison to edge computing. In the West, the tech industry is highly commercialized, and scalability means profitability. In contrast, given that the tech landscape is state dominated in China, it is less concerned with profitability and scalability. Tech in China is deployed and applied more often than not on a case-by-case basis with less malleability. In addition, as edge devices have lots of autonomy, edge computing inherently has fewer points of failure. This caters to the conservative attitudes of the Chinese. In essence, edge computing is more hardware based whereas fog computing is more software based. China is leading in hardware while software is more of Silicon Valley’s strong suit.
The Internet of Things is already a reality. Just look at Amazon Echo! However, in order to improve it and make it part of everyday life, lots of work is required to go into data processing systems such as edge and fog computing. With consistent efforts and funding, we should see these new technologies’ exciting applications in no time. Then we will truly feel like living on the EDGE!
*In this article, the West refers to mostly Silicon Valley in the United States, but also other western countries like Spain.
(Top photo from Pexels.com)