Johannesburg, 25 May 2021
Introduction
Edge computing is rightly being hailed as the cloud’s silver lining. It is a force multiplier which has triggered a step change in the performance and scope of the abilities of distributed sensing, computing and action. In short, it brings significantly enhanced benefits to users.
It is rare for a single emerging technology to have a truly disruptive impact by itself. Instead, it is typically the dynamic combination of a number of technologies which not only enables current products and applications to be “cheaper, faster and better”, but also opens new markets by enabling entirely new types of applications. Edge computing has emerged as such a key technology within a cluster which includes the cloud, the Internet of things (IOT), sensors and smart devices, 5G and artificial intelligence.
Edge computing has already found applications in a wide spectrum of sectors and industries, which include smart devices and appliances, banking(1), augmented reality(2, 3), electrical and water utilities(4), supply chains(5), drone security(6), geospatial applications(7), healthcare(8), real estate(9), manufacturing(10), oil and gas(11), fleet management(12), hyper converged infrastructure software(13), space(14) as well as smart buildings and cities(15,16) – to name but a few.
The emergence of the cloud
The Internet has been a mega-enabler which provides a location and device agnostic platform for electronically connecting people to people as well as people to computers. We are now being immersed in a world where the Internet is rapidly evolving into an extended platform which additionally connects smart devices to interact with one another autonomously without human intervention.
The maturing of the Internet and advances in computing and telecoms technologies inevitably led to the emergence of “the cloud”. It is a magical virtual place in cyber space where OPM (“other people’s machines”) store vast amounts of users’ data and provide a staggering amount of computing power, aptly referred to as “cloud computing”. From a user’s viewpoint, there is no need to even know where the “computers in the cloud” are, as they are accessed with Internet protocols. Neither is there any longer a need to own, operate or maintain expensive computing resources onsite anymore. All that is required is a cyber safe and reliable Internet connection. It is no surprise then that the cloud has become one of the cornerstones of modern enterprise IT architectures and that cloud migration is a key IT strategy imperative.
The cloud is now a de rigueur go-to technology for organisations and systems, which are intrinsically distributed in the sense that they consist of many remote connections that need to continuously interact with one another. These can include, for example, the collaboration of people who are not co-located, but instead working remotely from all over the world – a trend accelerated by the COVID pandemic. However, many systems increasingly also require remote devices to gather information (“sense the environment”) and share that information. In addition, many of the remote devices need to “take action”, underpinned by a decision regime. In contrast to standalone devices, their autonomous capabilities are synchronised with information which is shared by or act on instructions received from other remote devices or people.
The evolutionary blend of technologies which include artificial intelligence, 5G, robotics, advanced sensors and increasingly higher performing computing capabilities combined with the cloud have spawned the emergence of “smart devices”, which can do all of this.
The Internet of things (IOT)
The world first recognised the impact of “smart devices” with the introduction of smart phones. However, “smart” is such a powerful concept that it rapidly spread to all domains of technological application – and beyond. It seems that we now live in a world of “smart everything”, including smart devices and appliances, smart homes and buildings, smart cities, smart infrastructure and smart factories. It is almost a given that a smart device is networked and must be able to communicate, typically via the Internet/cloud.
It is evident that the IOT market is huge and growing. Gartner estimated that 8.4 billion devices were connected to the Internet in 2017, an increase of 31% from 2016.(17) Cisco estimates there will be 27.1 billion networked devices in 2021 (of which 43% will be mobile connected), a 58% increase from 17.1 billion in 2016.(18) According to Statista, the global IOT market is predicted to reach $520 billion in 2022. Other sources estimate an even higher number, with suggestions of 35 billion in 2020, and a growth of 152 200 connected IOT devices every minute by 2025.(19) Some estimate that the IOT device market will reach $1.4 trillion by 2027.
Walmart has indicated that it will manage more than 7 million unique data points in 2021. Gartner reported that the retail and wholesale trade segment included 440 million IOT endpoints in 2020, up from 360 million in 2019 and 290 million in 2018.
The spectacular growth in the number of IOT-connected devices, particularly smart devices, has inevitably also led to an equally spectacular increase in the amount of data being generated and transmitted across the Internet via the cloud. A self-driving car, for example, can easily generate six gigabytes of data per km travelled. Much of the data is just to indicate that “all is well”, however, and need not be continuously transmitted. This “chattiness” of smart devices is contributing to a data traffic jam on the Internet, which in turn degrades the performance for all. The IDC predicts that data generated by connected IOT devices globally will grow from 17.3 zettabytes (ZB) in 2019 to 73.1ZB by 2025.(20) This saturation comes at the same time when the “need for speed” in the cloud is increasing.
It seems that the cloud has become a victim of its own success. The cloud environment became “ripe for disruption”, begging for the intervention of another emerging technologies to not only resolve the dilemma but also to bring new applications and opportunities; mindful that technological disruption very often also spells the demise of mature technologies which are being rendered obsolete in the process.
Enter edge computing
The quest to reduce network traffic in the cloud fuelled innovation to develop smart devices which could place autonomous intelligence where it is needed, where the data is generated and where the action is required – at “the edge”. Hence the birth of edge computing.
In essence, edge computing refers to a system where smart devices with the ability to autonomously sense, compute, communicate and act, are connected to a network with the intention of gathering, processing and acting on local information, while simultaneously being informed by and informing the broader network with relevant information.
Advances in artificial intelligence, sensors, telecommunications technologies such as 5G, computing hardware and the maturing of blockchain technology – in essence the infrastructure to support the edge – enabled a new generation of smart devices, which could operate autonomously without the need to continuously revert to a central intelligent computer and controller, while communicating only the essential data required by other connections and devices.
The edge can be juxtaposed against a traditional cloud model, where the computing intelligence and storage of data are highly concentrated in a number of centralised computers located in a handful of hyperscale data centres which do the heavy lifting.(21) In order to be processed or stored, however, the vast amounts of data must first be transmitted via the Internet to the central cloud computers from wherever in the world it is generated, be it by people, other computers or devices. Needless to say, the exponential growth in data and its transmission in the cloud, has led to a saturation of the Internet creating the data traffic jam referred to above. One of the innovation drivers of edge computing was to relieve the bottlenecks that were beginning to throttle the advances that the cloud brought in the first place.
Edge computing has not only become an important technology in its own right, particularly with regard to addressing local issues “at the edge”, but also a key enabling technology in the cloud ecosystem. Rather than competing with the cloud per se, it complements the cloud in a synergistic manner.(22) The Linux Foundation's general manager of networking predicts that edge computing will overtake cloud computing by 2025, referring to the ability to compute and storage resources that are five to 20 milliseconds away.(23)
In addition to relieving the data traffic congestion on the Internet itself, which in turn enhances the performances of other cloud users, edge computing also brings a number of other benefits. One of the most important of these is the reduction of latency.(24) Simply put, latency is the time delay between “pushing the button” and waiting for the signal to arrive and subsequent response at another remote location.
In order for real-time systems, such as driving a car, for example, to be able to be operated remotely, extremely short latency times are required. Until recently, this was something that the technology-of-the-day was unable to deliver. The problem was exacerbated by the congestion-related delays in the cloud referred to above. The very short latency delays that have now become possible with the advent of 5G and edge computing have enabled a new generation of remote real-time experiences, ranging from remotely driven vehicles to enhanced augmented reality.
It is not difficult to understand why the force-multiplier impact of edge computing as an enabling extension of the cloud, in conjunction with its supporting infrastructure, is considered to be one of the most important contemporary emerging technologies.
As is often the case with disruptive emerging technologies, edge computing is also spawning business model innovation – the ultimate engine powering the wave of creative destruction that is at the heart of technological innovation.(25) Aruba, a subsidiary of Hewlett Packard Enterprises (HPE), predicts that new edge-related applications will include new revenue streams in the form of personalised experiences for shoppers, improved mobile experiences with turn-by-turn navigation or augmented reality, better business agility based on new availability of real-time data and less downtime and lower maintenance costs due to improved monitoring.(26) An HPE vice-president estimates that as much as 80% of IT, operations technology (OT) and consumer data will be generated at the edge, with most of it coming from IOT devices, emphasising the convergence of IT and OT (operations technology) as well as “edge as a service”.(27) The vice-president of digital transformation for Rockwell Automation predicts that “real-time availability of mission-critical workloads will be vital for companies scaling smart factory initiatives in 2021”.(28)
A recent report from Forrester suggests that 2021 will be the year when “this emerging technology [edge computing] graduates from experiment to practically applicable tech, driven largely by AI and 5G”.(29,30) The performance advantages of 5G now being rolled out, including private 5G networks, enables multi-layer “onion-skin” solutions. An edge layer can connect with an intermediary “meta edge computing” layer, sometimes referred to as “fog computing”, which in turn connects to the broader cloud.(31,32) Heavy duty computing power which may not be feasible (yet) to be accommodated in individual edge devices can be located in the intermediary layer, where a 5G connection with the edge device will still have an acceptably low latency.(33) Technological advances have now enabled complex and large AI algorithms which could previously only be accessed onsite or via cloud computing, to be edge-based, simultaneously addressing a number of the shortcomings of cloud-related AI.(34,35,36) The research director in the IDC’s worldwide infrastructure practice focusing on edge strategies, notes: “AI is the most common workload… there has been an increased interest in applying AI at the point of generation for real-time event detection.”(37)
“Edge computing is the smart silver lining of the cloud. Technological advances now enable us to place smart IOT devices and sensors "close to the action". Not only does this provide computing power where it is needed, it also significantly reduces latency and network traffic in the cloud. Edge computing brings the potential of an entirely new range of applications, which will disrupt many current business models," says Sivi Moodley, CEO at Macrocomm Group.
Edge adoption
The IDC estimates that by 2023, more than 50% of new enterprise IT infrastructure deployed will be at the edge rather than in corporate data centres, up from less than 10% in late 2020(38); with the number of apps at the edge increasing 800% by 2024. It also predicts that the edge computing market worldwide will grow to $250.6 billion by 2024.(39) Gartner estimates that whereas less than 10% of enterprise-generated data is currently created and processed at the edge, this will grow to 75% by 2025.(40)
McKinsey suggests that the edge computing market will represent a value of up to $215 billion in hardware by 2025.(41) The non-profit organisation State of the Edge estimates that the edge computing infrastructure market will be worth $700 billion by 2028. Although not all edge devices use AI, it will increasingly become a major driver for edge computing. One estimate puts the revenue for AI-powered edge devices and networks deployed at $827.6 billion by 2025, up from $127.5 billion in 2019. Deloitte predicted that more than 750 million edge AI chips (specifically designed to perform or accelerate on-device machine learning) would be sold in 2020.(42)
Keeping the edge safe
Since edge computing reduces network traffic, it inherently reduces cyber risks. Even so, cloud and edge computing remain vulnerable to the scourge of cyber-attacks and intrusions, as is the case with networked IT configurations in general. Already in 2018, it was estimated that the average amount of time that it takes for an IOT device to be attacked once connected to the Internet was only five minutes.(43) A range of dedicated cloud-based cyber security measures are continuously being developed and improved to protect the cloud and the edge, including secure access service edge (SASE).(44) Gartner predicts that more than 40% of enterprises will have plans to adopt SASE by 2024, an increase from less than 1% in late 2018.
The way forward
Edge computing is a powerful technology that contributes significantly towards enhancing the scope and performance of cloud-based applications. It provides game-changing benefits to users, particularly when combined with artificial intelligence. Edge computing is a young technology and hence one can optimistically anticipate increasingly higher performance, reliability and more use cases, which are typically also accompanied by falling prices. The growth predictions for the edge computing market is encouraging.
[1] Deltec bank, Bahamas explains edge computing and its impact on banking, AB News Wire, 13 Apr 2020
[2] How augmented reality is giving business apps an edge over competitors, Business Times Online, 16 May 2020
[3] Demystifying edge computing, Forbes, 15 Sep 2017
[4] Embedding AI-powered analytics into smart meters at the grid edge, IoT.do, 18 Feb 2020
[5] Edge computing meets 5G: Rethinking the intelligent supply chain, Manufacturing, 23 Jan 2019
[6] Dedrone secures the world’s airspace with distributed edge computing, sUAS News, 3 Dec 2020
[7] How artificial intelligence can enhance GEOINT at the edge, C4ISRNet, 9 Oct 2019
[8] GE Healthcare unveils new edge computing tools for clinicians, Healthcare IT News, 16 Oct 2020
[9] How edge computing may revamp, revitalize commercial real estate, ZD Net, 3 Jan 2020
[10]IIoT: Chemical makers approach the edge, Chemical Processing, Mar 2020
[11] Saudi Aramco, Stratus and Aveva to discuss edge computing in oil and gas digital transformation, Arabian Industry, Nov 2020
[12] How to explain edge computing in plain English, The Enterprisers Project, 30 Nov 2020
[13] Evaluate hyperconverged infrastructure software choices for edge computing, Gartner, 22 Feb 2021
[14] HP enterprises will give edge computing some space on ISS, Interesting Engineering, 14 Feb 2021
[15] Cities are combining edge computing and IoT data collection to improve services and save money, Tech Republic, 19 May 2020
[16] Smart cities look to the edge for next level urban planning, Real-time Insights, 2 Jul 2020
[17] Computing at the edge: Why IoT requires decentralization, Crypto Briefing, 5 Jul 2019
[18] How will IoT technology be applied in 2021 and beyond? IoT trends to watch, NEC, 17 Feb 2021
[19] Re-Hashed: 27 Surprising IoT Statistics You Don’t Already Know, The SSL Store, 17 Feb 2021
[20] A zettabyte is 1,125,899,910,000,000 megabytes.
[21] 4 things you need to understand about edge computing, Venture Beat, 29 Mar 2020
[22] Multi-mode interaction amongst technologies, Research Policy, 1997
[23] Linux foundation exec believes edge computing will be more important than cloud computing, ZD Net, 23 Sep 2019
[24] Centurylink targets 5 milliseconds latency with edge computing investment, RCR Wireless, 13 Aug 2019
[25] 5 edge computing predictions for 2021, Tech Republic, 27 Oct 2020
[26] Harnessing the power of the edge, Aruba, 2020
[27] 2 IoT technology trends reshape the edge (Source: TechTarget, 21 Apr 2020), TechTarget, 21 Apr 2020
[28] Will edge computing become the new cloud in 2021?, Tech Republic, 25 Nov 2020
[29] Edge AI in a 5G world – part 2: Why make cell towers smart?, Ubuntu, 2 Apr 2020
[30] 5 edge computing predictions for 2021, Tech Republic, 27 Oct 2020
[31] Comparing edge computing vs. cloud computing, Internet of Things Agenda TechTarget, 2 Oct 2020
[32] Fog computing vs edge computing, Erpin News, 19 Jan 2018
[33] Edge AI in a 5G world – part 2: Why make cell towers smart?, Ubuntu, 2 Apr 2020
[34] On how edge computing and blockchain will be key in different ways for AI, The Internet of Things, 10 Jun 2019
[35] How to explain edge computing in plain English, The Enterprisers Project, 30 Nov 2020
[36] Pushing computing to the edge by rethinking microchips' design, Eurasia Review, 28 Feb 2021
[37] Edge computing and AI: 7 things to know, The Enterprisers Project, 5 May 2020
[38] How to explain edge computing in plain English, The Enterprisers Project, 30 Nov 2020
[39] Will edge computing become the new cloud in 2021?, Tech Republic, 25 Nov 2020
[40] How to explain edge computing in plain English, The Enterprisers Project, 30 Nov 2020
[41] State of the Edge report for 2020, Connected World, 25 Dec 2019
[42] Edge computing and AI: 7 things to know, The Enterprisers Project, 5 May 2020
[43] Re-Hashed: 27 Surprising IoT Statistics You Don’t Already Know, The SSL Store, 17 Feb 2021
[44] Will edge computing become the new cloud in 2021?, Tech Republic, 25 Nov 2020
Share