Edge computing will replace cloud computing?

Not so long ago, my colleagues and I discussed the dynamic development of technologies and the prospects for the development of cloud computing. And just stumbled upon an article.

Each new technology replaces the old one. Sometimes, as in the case of the cloud, the old technologies are re-branded to make them more attractive to consumers and, thus, create the illusion of a new product. Cloud computing previously existed in one form or another. At one stage, they were called “on demand computing” (computer resources on demand), and then transformed into an “application service provider” (ASP).

Now there is edge computing (note there is no well-established Russian term, but it can be translated as "the concept of boundary computing" or "peripheral computing") . to which industry observers and experts predict the ability to replace the cloud. But the question is: will it really happen? After all, the same thing was said about radio when television was invented. Nevertheless, millions of people turn on the radios every day, but it has changed the format, now it’s mostly listened to in the car.

Of course, there are some technologies that really carry out a revolution in changing people's habits and their way of thinking. People used to listen to music from Sony Walkmans, and today we use a smartphone everywhere.
')

Levin's Prophecy

So why do people think that edge computing will conquer the cloud? This statement has been stated in many articles. For example, Clint Boulton in March of this year writes about this in the article "edge computing replaces the cloud." He refers to venture capitalist Andrew Levine, general partner of Andreessen Horowitz, who believes that more computing resources will move in the direction of end devices - such as unmanned vehicles and drones - that make up at least part of the Internet of things. Levin predicts that this will mean that the cloud has come to an end, because data processing will move back to edge computing.

In other words, now there is a tendency to centralize computing in data centers, while in the past they were often decentralized or located closer to the place of use. Levin sees the unmanned vehicle as a data center: they have more than 200 processors capable of providing full fault tolerance, so as not to cause an accident on the road. The nature of autonomous vehicles means that their computing power must be independent and in order to provide security, you need to minimize any communication that they have with the cloud. However, they cannot completely do without him.

Complementary models

These two approaches can complement each other. Some of the arguments for edge computing simply fall away when it comes to increasing data volumes, which lead to an even more depressing and slow network. The delay is the culprit. Data is becoming more and more: the amount of data per transaction, "heavy" video and a lot of data from different sensors are increasing. Virtual and augmented reality will play an increasing role in its growth. With such an increase in data volume, solving latency problems is more difficult than it was before. Now it makes sense to place the data closer to unmanned vehicle devices in order to eliminate the delay, but nevertheless most of the data is still located remotely in the cloud. The cloud will continue to be used as a provider of services, such as media and entertainment. It can also be used to back up data and to exchange data from a vehicle.

A bit away from autonomous vehicles and back to the more familiar business. Creating a number of small data centers or disaster recovery sites can reduce the effect of scale, as a consequence, increase costs and make work less efficient. Yes, the delay can be reduced, but in the event of a catastrophe, the consequences will be no less disastrous; therefore, to ensure business continuity, some data should be stored and processed elsewhere — in the cloud. In the case of unmanned vehicles, in particular, because they must work regardless of whether there is a network connection or not, it makes sense that certain types of calculations and analysis are performed by the vehicle itself. However, this data will still be backed up to the cloud when the connection is available. The approach will be hybrid: edge and cloud computing will complement each other, and not be used alone.

From the periphery to the cloud

Sage Scaria, senior director of consulting firm TCS, offers several examples where edge computing can be useful. In his article on LinkedIn Pulse, “Edge computing vs. Cloud computing: for whom is the future? He does not believe that the cloud will be finished.

"Edge computing will not replace cloud computing ... in fact, an analytical model or rules can be created in the cloud and then used by end devices ... and some [of them] are capable of analyzing." He then goes on to talk about fog computing, which involves processing data from the periphery to the cloud. He believes that people should not forget about data warehouses, as they are used for "slow analytical queries and massive data storage."

Edge will win the cloud

Despite this argument, Gartner analyst Thomas Bitman believes that “Edge computing will“ eat ”the cloud.” “Today, cloud computing is devouring enterprise data centers, more and more of the burden is on the cloud, and some are being transformed and moving to the cloud ... but there is another trend that will shift workloads, data and business value away from the cloud ... And the trend towards edge computing is even more important and powerful than the trend of cloud computing.

Later in his blog, Bitman writes: “The speed of cloud solutions alone is not enough. Massive centralization, economies of scale, self-service and full automation overcomes only halfway through - but all this will not overcome physics - the weight of data, the speed of light. How people should interact with digital reality in real time, if there is a delay from the data processing center located thousands of kilometers away. Delay is important. I exist here and right now. Play the right advertisement, before I turned away, point to the store I’m looking for when I drive, help my unmanned car choose the right path through a busy intersection. And all this I need to get NOW. ”

Data acceleration

Bitman makes some fair comments, but he uses an argument that is often used in relation to delays and data centers: they must be located close to each other. The truth is that global networks will always be the foundation of edge computing and cloud computing. Secondly, Bittman clearly did not come across data acceleration tools such as PORTrockIT and WANrockIT. While physics is certainly a limiting and complex factor that will always occur in all kinds of networks - including WANs, today you can place your data centers at a distance from each other. The delay can be reduced, and its impact can be leveled, regardless of where the data is processed, and no matter where the data is stored.

Therefore, edge computing is not a new breakthrough solution. This is just one solution, just like the cloud. Together, these two technologies can support each other. The difference between edge computing and cloud computing is that "edge is a method for accelerating and improving cloud computing performance for mobile users." Thus, the argument that edge computing replaces cloud computing is very unconvincing. For marketing reasons, cloud computing may be renamed, but the essence will remain the same.