We live in times where every other new innovation in the technology space is touted as the equivalent of “death” of an existing technology.
Amid all the lack of complete understanding and marketing hullabaloo, the mere possibility of the “old” and the “new” working in tandem is ignored. Then, there is the very interesting case of technologies that are packaged as revolutionary, when they should be more correctly described as better versions of existing technologies.
Let’s remember that to some extent, even cloud computing was heavily marketed to create the illusion of something revolutionary, when the same technology had existed, albeit not as pervasive as the cloud became, as on-demand computing, and application service provisioning. Where does “edge computing” stand, then? First, by understanding edge computing we will get a better idea of how to compares to cloud computing. Let’s explore.
Basic question of technological disruptions
Industry commentators seem to be very clear – edge computing will replace cloud computing. We ask – will it, really? Radio was supposed to be dead soon after televisions hit the scene. Everybody knows people who frequently tune in to radio, even today. The radio is critical for anyone who drives a car. You should not be watching TV and driving at the same time – you would be a public menace if you were trying to pull that off. Sort of like Det. McNulty was in “The Wire” when he chose to drive drunk! But he was a dam good detective!
Then, of course, there have been technologies that have replaced others. For instance, smartphones have obliterated pure MP3 players to a great extent, with advanced music playback options. Back to our question for now: let’s understand the origins of “edge” will replace “cloud” argument.
Origins of ‘cloud will make way for edge’ idea
The idea of edge computing’s potential to replace cloud computing first popped up when Clint Boulton raised it in Asia Cloud Forum, in an article called “Edge Computing Will Blow Away the Cloud” in March of 2017. He cited Andrew Levine, a venture capitalist, in particular. Levine strongly opines that data processing and computational resources will move toward “edge computing.”
A lot of devices such as drones and self-driving cars require the kind of data processing that is best delivered by edge computing.
How does IoT accelerate the onset of edge computing?
Until now, the market has witnessed a trend to centralize computing within the datacenter. Years ago, computing was localized at the point of use. Self-driving cars, for instance, leverage dozens of CPUs that work in tandem to keep the vehicle safely on road. To enable such autonomous operations of technology, systems should ideally minimize all kinds of reliance on the cloud. Such devices are the backbone of IoT, which is a very real transformational force in play.
Understanding the real need for edge computing
The IoT is a real force, and it’s set to be a major part of the cloud. The emergence of such massive numbers of Internet-connected devices implies serious data processing challenges for cloud machinery. Cisco has predicted that cloud traffic will increase to four times its current volume by 2020. Now, IoT is all about meaningful processing of device generated data, and the cloud is about processing data obtained from centralized computing.
Simultaneous management of stratospheric growth rates in data, hence, is a major issue. Edge computing helps IT meet this challenge. This is how the idea of edge computing as a necessity that will replace cloud computing is placed.
Understanding edge computing with a deep dive into the cloud
Computing at the edge helps keep the data processing close to the data source. Then, processes are divided between the central (such as AWS, Google Cloud, Azure) and the edge system. Some regard it as a cloud/server architecture, with additional decision-making capability around what to process at the client level, and what to process at the server level.
So, for IoT and similar highly distributed applications, we have a client/network edge/server architecture, and for devices without their own processing power, we have a network edge/server architecture in place. For applications that need super-short response times, centralized cloud makes way for edge computing. However, cloud computing remains as relevant as ever, for applications that don’t have excessive demands on response times.
Case for edge and cloud computing working in tandem
Take it from us, edge computing and cloud computing are entirely different. There’s just too much confusion around the concepts, which creates scary questions such as “how soon can edge computing replace cloud,” and “what do I do with my investments in the cloud?” The global IT thought leaders and innovators should think on the lines of creating purpose-built applications based on edge computing.
‘Edge vs. cloud’ or ‘edge of cloud’?
A salient example would be a sensor in a drone that should be able to quickly do the computations (edge computing) and make course corrections for accurate air-travel. The question everyone should ponder is – what percentage of the total global computation workloads will require the kind of near real-time response times that edge computing delivers?
Cloud computing will continue to be relevant for the bigger chunk of workloads, right from enterprise data analytics to resource planning. What’s more, even a hypothetical scenario where such massive data processing is moved at the edge will result in unimaginable and unmanageable security, privacy, and management issues.
Response from public cloud
During times when panic-mongering commentators are busy waiting for the day when the globe’s cloud computing magically moves to an edge computing model, there’s some real work being done by public cloud innovators and thought leaders. Almost all major public cloud service providers have strategies in place to manage IoT’s data processing needs, and technology stacks that will deliver edge computing.
Of course, edge computing will work for applications with special needs, and cloud will continue to be the general purpose platform that can also work seamlessly with applications running on edge computing.
Data volumes are burgeoning. Applications need near real-time response to data processing requests. Latency is the common denominator for several transformational forces of technology, such as IoT, artificial intelligence (AI), and virtual reality (VR). Edge computing, indeed, is a solution to all these applications. By understanding edge computing, you can clearly see it will complement the all-pervasive cloud computing and create a hybrid solution that will be absolutely constructive.