Edge computing is the sinking and extension of cloud computing capabilities to the edge and user sides

This paper analyzes the underlying logic of edge computing from the perspective of revenue logic , Edge computing is the sinking and extension of cloud computing capabilities to the edge side and the user side , Solve the wide link problem that cloud computing cannot meet through cloud edge collaboration 、 Low latency 、 Good control 、 Low cost user needs .
edge , It is the place where computers touch the world , From this big data 、 Artificial intelligence is no longer superior , Instead, it is actually in the factory 、 coal 、 oil well 、 The distribution network and other sites bring value into play 、 An agent that generates revenue .

About edge calculation , It is often compared to the cerebellum of an octopus in the industry ： Octopus are very quick and agile in hunting , There's a great fit between the arms and feet , Never tangle and knot . This is due to their similar distributed computing “ Multiple cerebellums + A brain ”, And many of the cerebellums are like edge computing , Through the eight tentacles and the perception of the outside world 、 In situ calculation to achieve the effect of rapid processing of complex problems .

5G And Internet of things development , Bring explosive growth of data , And higher requirements for computational power . Low delay 、 Low bandwidth consumption 、 Internal requirements such as wide links make edge computing imminent .2018、2019、2020 Three years in a row , Edge computing is the largest in the world IT Consultancy, Gartner named “ Ten strategic technologies ”, According to the Gartner To measure , Is expected in the future 3-5 year , Edge computing will become the next blue ocean market with a direct market size of more than trillion .
From the above news and introduction, we can basically draw a conclusion ： Edge computing is booming , At present, it is in the expansion period of technology maturity expectation ！ The sound development of a technology ultimately needs to conform to the business logic , The author starts the discussion from the revenue logic behind the edge calculation , I hope it can bring some thoughts to the enterprises in the process of digital transformation .
Case a
A technology company is a high-tech enterprise in the software and information technology service industry , The gross profit margin has exceeded for four consecutive years 90%, High gross profit margin plus high net profit margin , The earning power of the company belongs to the super type .2016 year -2019 The annual gross profit rate of the company is respectively 92.30%、93.66%、94.29% and 93.27%. This is because licensed algorithm software packages generally do not require the production of physical hardware 、 Packaging and transportation , Therefore, the main business cost is low , Higher gross margin .
Case 2
Someone began to count from zhaokuangyin's yellow robe , Earn... Every day 208 ten thousand （1 Shuang =6.4 Billion ,77 One day's income 1.6 Billion , Average daily income 208 ten thousand ）, until 2021 year , Earned a total of 800 Billion .《 Forbes 》 The data shows that Bill · Gates is now the fourth richest man in the world , Is worth more than 1300 Billion dollars , renminbi 8414 One hundred million yuan , Conclusion ： Still no bill · Gates has money ！

There is a logic problem behind the above two cases ： Copy income by value , It will be faster . A technology company authorizes its own face recognition algorithms “ sell to ” Downstream channel integrators 、 Smart device manufacturers make a profit , But the product of algorithm only needs some investment in the design stage , And as the number of promotion increases , The marginal cost is almost 0！
bill · Gates passed Wintel The alliance will use its own operating system 、 Office software and other mature software achieve the same purpose through horizontal replication . This is also the truth behind it , That is, the ordinary income model depends on ability to make money ,Buff The model is to make money by value , A summary chart is as follows ：

This is the underlying revenue logic , By ability VS Driven by the return between values over time 、 Changes in the size and structure of the organization can only be more and more different , Yes, of course , There is nothing wrong with earning money by ability , After all, it's safer .
Under such revenue logic , For example, in the product dependent mode , It puts forward some important characteristic requirements for the product itself ：
The first is the value dominance , It is better to open the box and have value , This requires the product to have a direct connection with the physical world , It is better to have interaction ;
The second is that the product is easy to scale , A good test is to assume that 1 A small target （ According to the folk saying 1 A small target =1 Billion ） To spend on this product , Can this product hold up ;
Then there is the pre-sale of this product 、 Market publicity , Is it easy to be accepted by customers ; The last is whether it can be delivered well , The best thing is Ctrl+V The pattern of . To sum up, the product needs to have ： Value dominance 、 It can be big or small. 、 Try before you buy and have strong reproducibility . So from the market 、 sales 、 pre-sale 、 deliver 、 You can get through the chain of operation and maintenance , Coordinate the interests of all parties , Through such a product , Stir up some splashes of the times .
Combine these characteristics with edge computing technology , It can lead to the following thoughts ：

Value dominance

edge , It is the place where computers touch the world , It was born on the spot , It is closely related to the production business of customers , The traditional centralized processing （ Such as cloud computing ） Too far away from the site , There may be a problem that the value is not directly reflected . From this point of view , Edge computing inherently requires the algorithm model running on the edge side to be value explicit , Otherwise, it will be eliminated ！ For example, the precise dosing system of Merrill Lynch data in the water sector , Help typical customers save after investment 20% About the dosage , On a daily basis 40 Take a water purification plant with a scale of 10000 tons of water purification as an example , It can help customers save 560 About ten thousand yuan for medicine .

It can be big or small.

Back to 1 A small goal , Unless it's a huge system , For example, large rockets 、 Big planes, etc , Traditional software vendors want to 1 A product is worth 1 A small goal is hard to imagine . And edge computing is naturally characterized by horizontal scalability in actual business scenarios , For example, the water treatment process of waterworks is not bad 、 The process of sewage treatment in sewage treatment plant is basically based on biochemical reaction 、 The basic principle of the thermal power plant is almost the same , At this time, edge computing can be deployed in different factories 、 Different sites, etc , Because the expansion itself is only limited by the scale of the business , In principle, there is no upper limit for such investment , This is the biggest attraction of edge computing .
For example, the precise aeration system of Merrill Lynch data for sewage plants , The optimal space for dissolved oxygen concentration of the sewage plant after operation is 22% about , The constructed precise aeration model saves the average aeration amount 30% about , Directly help customers to save electricity bills every year 500 All around , Save labor cost 30 ten thousand , For the group model , On the one hand, through cloud side collaboration, customers' ideas of pilot first and then promotion are met , It can be minimized and deployed in a sewage plant , In the later stage, the cloud edge collaboration platform will be promoted to directly cover the whole group , Each sewage treatment plant is an edge end , The headquarters of the group is the cloud .

Try Before You Buy

In tradition IT During the pre-sales process , The general communication is PPT And similar materials , This kind of material often resonates with the personnel of the information department , But for the business unit that finally decides to pay the bill , In particular, the production business side is often insensitive . Business departments often pay more attention to whether they can reduce costs and increase efficiency when purchasing products 、 Expand revenue and other issues , Software products are generally deployed centrally , It's hard to interact directly with the physical world , It makes it more difficult to sell before sales .
This problem can be solved by means of edge calculation . With the help of the feature that the edge calculation can be large or small , Minimize deployment to make product value explicit . For example, deploy a Merrill Lynch data operation site security control system , You can get the live video data directly through the camera , Use edge calculation analysis , Directly analyze the violation results （ If the safety helmet is not worn 、 smoking 、 Face recognition results, etc ） Send to the cloud for Tempo BI Visual display , Business department personnel can directly immerse themselves in , When you say nothing at all .

Highly reproducible

Edge computing is mainly around the regular model 、 Mechanism model 、 Artificial intelligence model, etc , Not design , The design phase of a good algorithm model often relies on capability delivery , But in the reasoning stage, we can rely on edge computing to achieve value delivery , It is precisely because this kind of algorithm model is highly reproducible , Implementation cost of investment 、 Development costs 、 The cost of operation and maintenance is relatively low , Then delivery is relatively easy .
such as Merrill Lynch data Deeply cultivated in the wind power field for many years , Precipitation of many mature algorithm models in the field of wind power , For example, the gearbox fault early warning model 、 Early warning model of unit performance degradation 、 Fan fault diagnosis mechanism model, etc , These applications can meet the fault diagnosis of fans in different wind farms 、 The internal demands of refined operation .
Merrill Lynch data intelligent cloud side collaboration platform is such a powerful tool and platform , It has accumulated a lot , Bearing the Tempo AI And many data analysts 、 Algorithm Engineers year after year 、 Many industry-level algorithm models that are abstractly summarized day after day in combination with customer business , Help customers successfully carry out enterprise digital transformation , Better release value , You can say that , Edge computing is a powerful means to connect the last mile of enterprise digital transformation , It is connected to the enterprise application mall , Butt joint the site SCADA、PLC etc. , Breaking the bandwidth limit , Reduce on-site business operation delay , Silently guarding the more optimized operation of the production site , Constantly create value for the enterprise .

edge , Where computers touch the world . In this technology feast that directly benchmarked with Cloud Computing , Both upstream and downstream of the industrial chain are taking positive actions , Strive to stride from the industry Intelligent Cloud era to the edge era , Edge computing will permeate thousands of industries like hydropower , Service demand will flourish . Merrill Lynch data will always hold its head high , Down to earth service for enterprise digital transformation , Look up at the sky and stand at the forefront of technology in the great era , future , Has come ！