Understanding the concept of cloud computing from an economic perspective

Cloud computing initial fixed investment high, marginal input low, which leads to diminishing marginal cost, marginal income increment. In cloud computing, both infrastructure, platform and software require higher initial fixed inputs. Once built, however, it can be reused over and over again, and each value-added business, with a lower marginal input, can be launched without developing infrastructure, platforms and software from scratch.

What is cloud computing? Now is not without answers, but too many answers.

Now the answer to "What is cloud computing" is so much that people can't figure out what cloud computing is.

The reason is that people just inductive local phenomenon, just grasp the cloud computing this elephant a part of the cloud as a whole.

Of course, a comprehensive induction, than the local induction is not much better, it makes the cloud computing infinitely complex, into details, still confused what is cloud computing. Like the whole of the elephant cut into countless fragments, yards into a piece of the body, so that people still do not see what is.

Therefore, we need to make a theoretical generalization, the most essential kernel of cloud computing, so that this kernel is not submerged in its phenomenon extension.

Here, I intend to do a core analysis of cloud computing from the very essence of economic theory. So that people who do not understand the technology can also directly grasp its commercial essence.

I think the core feature of cloud computing is centralization and decentralization to grasp resources. From the phenomenon, on the one hand, in the cloud to share the digital resources, on the other hand, in the end to improve the efficiency of distributed computing; its economic implication is that the initial fixed input and marginal input between the Division of labor, through large-scale sharing of the former, safeguarding the latter's customization, overall to achieve the effect

Perhaps the text we use to interpret the kernel is not reduced by the simplicity of the kernel, but at least we can do it around a central center rather than taking cloud computing as a basket and putting anything irrelevant into it.

The kernel of cloud computing from the experience

Before entering the economic theory analysis, we first from the experience, check our proposed cloud computing the simple kernel, whether conforms to the technical sense cloud computing.

Cloud Computing's core can be summed up as "a center, two basic points", a central, is to improve efficiency (that is, pay increase), two basic points, one is to share, and the other is to disperse value-added. If you can say one more word, it is: in the focus of sharing and decentralized value-added, through the interface (such as the API) to join.

A good definition should be concise enough without omission. We can check this definition.

Google CEO Eric. Mr Schmidt argues that cloud computing, unlike traditional pc-based computing, distributes computing and data across a wide range of distributed computers, making computing power and storage highly scalable and facilitating access to the network for Applications and services through a variety of access modes.

This definition takes into account the relationship between centralization and decentralization, especially the characteristics of distributed computing, and the connection method is mentioned. But there is no clear emphasis on the focus. In this respect, Kai-fu Lee has added "let the Internet" as a "cloud" to each Netizen's data center and Computing Centre "this feature." None of them mentioned user value added and on-demand access, although Google has personalized this in actual combat.

Merrill Lynch believes that cloud computing is delivering personal and commercial applications from centralized servers via the Internet. These servers share resources, and through sharing, resources can be used more efficiently.

In this definition, the focus and the sharing are mentioned, and the decentralized applications (individual applications and commercial applications) are also taken into account. However, it is not clear that the distributed characteristics of the application, do not mention the value-added of the application, did not explicitly mention the link.

The father of Grid computing Ian Foster that cloud computing is a model of large-scale distributed computing, driven by the economies of scale. In this mode, some abstract, virtualized, dynamically scalable and managed computing power, storage, platforms, and services are pooled into resource pools and paid to external users on demand via the Internet.

This definition is more comprehensive in that it refers to a centralized aspect ("massive") and also to distributed computing, as well as sharing ("virtualization", "resource pooling") and value-added ("on-demand delivery"). In particular, it refers to efficiency ("economies of scale", that is, the increase in scale returns, but ignores the increase in scope returns).

Second, the economics explanation of cloud computing

Looking at cloud computing from an economic rather than a technical point of view, there are two aspects that are different from the traditional economy. One is the increment of compensation, the second is the separation of fixed initial input and marginal input, which reflect the technical characteristics of the new economy different from the old ones.

1, the first major difference, cloud computing presents an increase in returns, not only the increase in scale, but also the scope of compensation increase.

1 The difference between the interpretation of neo-classical interpretation and the new economy: industrial economy and network economy, reflected in cloud computing

The neoclassical economic theory and the new economic growth theory, in a sense, represent the different interpretations of the industrial economy and the network economy to the world.

The efficiency assumption of neo-classical economic theory is that the compensation is decreasing or the remuneration is unchanged; the efficiency assumption of the new economic growth theory is the increment of remuneration. This is the fundamental difference between the two. This difference reveals the biggest difference between the industrial economy and the network economy.

First of all, it needs to be explained that the efficiency point of view, rather than the efficient view, is because the perspective of efficiency does not distinguish between cloud computing and traditional industrialized economic and technological characteristics. We often hear that when people talk about the benefits of cloud computing, they say that cloud computing can reduce costs, improve profitability, and improve efficiency. But that's not the point. Because almost any technology can reduce costs, improve profitability and improve efficiency, it cannot separate cloud computing from traditional industrial technology.

Efficiency is in the market size and scope of the unchanged conditions said. Effectiveness is the rate of change in efficiency, when the efficiency (cost and benefit ratio) in the size or scope of the change in terms of, industrial technology and network technology will present (tangent slope positive and negative signs) the opposite characteristics, such as can be seen in the same efficiency, the difference between the increase in returns and the decline in

There is a particular reason to do so for the economic interpretation of cloud computing. Only by adopting the perspective of efficiency can we reveal the economic effect of the cloud and the end. Because of the increase in remuneration and diminishing returns, to a large extent, it is determined by the relationship between initial fixed input and marginal input. The cloud corresponds to the initial fixed input, and the end corresponds to the marginal input.

The increase in returns is determined by the economic and technological characteristics of the initial fixed input and low marginal input. The economic and technical characteristics of material production in traditional economy are low initial fixed investment and high marginal investment, which leads to the increment of marginal cost and diminishing marginal income. It is hard to use shared infrastructure, platforms, and software to reduce marginal inputs, like cloud computing.

Cloud computing initial fixed investment high, marginal input low, which leads to diminishing marginal cost, marginal income increment. In cloud computing, both infrastructure, platform and software require higher initial fixed inputs. Once built, however, it can be reused over and over again, and each value-added business, with a lower marginal input, can be launched without developing infrastructure, platforms and software from scratch.

This is precisely the difference between neo-classical theory and new economic growth theory. Scholar Yong has described the difference between neo-classical theory and new growth theory: "neo-classical theory assumes that total production function has the same property of scale." The key amendment of new growth theorists to the neoclassical growth model is to consider technology factor A as an endogenous variable of economy, so that the total production function in the new growth model shows the property of increasing scale returns. ”

What do people figure in cloud computing? The figure is the increment of the reward (or increase in income). Taobao merchants do not have to build their own infrastructure, do not have to build a platform, do not have to develop their own software, only a small amount of marginal investment in their differentiated business, you can directly achieve economic results. Cloud computing as a technology factor, for the initial fixed input, the efficiency direction can be changed, so that the total production function is increasing the nature of the reward, because cloud computing has the positive correlation between scale or scope and cost additive (weak cost) in the transformation of technology to economy. Generally speaking, with one plus one more than two economic and technical characteristics. This is the opposite of traditional industrial technology.

Now we see the first step in using economics to explain cloud computing: instead of simply saying that cloud computing can reduce costs, improve profitability, and improve efficiency, it can be accurately expressed as a mathematical marginal concept, which translates into "more ... The more ... "the expression, for example, the greater the size, or the more complex the scope, the more the cloud can reduce costs, improve profitability, improve efficiency. The implication is that, in the case of scale and scope variables, if cloud computing is used, efficiency will be reduced from low to high; At the same time, it is also an objective indication of the conditions under which cloud computing--in the small size and scope of our country--is not working, and cannot be blindly adopted.

2 interpretation of two different directions within the new economy

The above mentioned, although can distinguish the cloud computing in the industrial economy and the network economy nature difference, but cannot differentiate railroad economy, the telecommunication economy and the Internet economy. Because the railroad economy, the telecommunication economy is also the network economy. Only by increasing the rate of return, it is still possible to confuse the cloud computing and the Internet in the sense of cloud computing. For example, there is no way to differentiate the nature of cloud computing in a telecoms-and-internet-mixed industry involving mobile internet. It is therefore further to lock down its uniqueness.

We can note that the same interpretation of the network economy, in the new economic growth theory, there is a common people even economists are unfamiliar with the bifurcation, can be used to explain the traditional network economy (atomic or analog signal network economy) and the digital network of the similarities and differences.

The above mentioned Yong said the new growth model, only related to the increase in the scale of income, in contrast, is the scope of income increase. Since Chamberlain's theory of incremental pay, the new growth theories have developed two opposite theories of returns, one of which is the dominant form of return on scale, such as Stigglitz, Krugman and Romer; and one is the theory of the Panzer of Northwestern University, which raises the scope of reward.

In today's opinion, the theory proposed 30 years ago, as if it were specifically tailored for cloud computing. Can just explain the special principles of cloud computing. Unlike telecoms economies of scale, there is a wide range of economies in cloud computing. The sharing of initial fixed inputs on the platform, such as API-open platforms and development tools, has resulted in a distinct variety of varieties. The application of it in mobile Internet is obviously more differentiated than the application of traffic technology.

For example, on Apple's iphone, Chor Kom 2008 developed a program with 26 different fart sounds to make 400,000 of dollars, driving the global Fart program development heat, forming economies of scale. Apple is outraged, in the App Store's iOS application review of the basic outline, in the metadata to the user interface and other technical standards, explicitly added a: "We do not need fart software." "Obviously, Apple launched the App Store platform, welcomed the diversification of product development, do not want everyone to rush to fart this single variety up." Since then really refused to continue to "defy the" new fart program pull my Finger. Apple is now on cloud computing, and he will feel as much as Google about the importance of cloud computing to the richness of the value-added business.

As a result, we can further elaborate the "one center" of cloud Computing as: cloud computing can improve performance, especially to improve the efficiency of diversification, but also to improve the efficiency of large-scale. If you have to express cloud computing with imprecise concepts such as cost, profitability, and efficiency, be sure to add a limit, for example, that cloud computing is good for saving costs, improving profitability, and improving efficiency in complex conditions, or foster: Cloud computing has "economies of scale". I used the market size and scope of two conditions, the economic role of the cloud to qualify, that is for this consideration.

2, the second major difference, cloud computing will be the initial fixed input and marginal input of the main body, in a complementary way to join.

In the traditional economy, centralized business and decentralized business are often the same enterprise, while in cloud computing, the focus of business and decentralized business of the main body, unexpectedly is different enterprises.

Under traditional conditions, such as American retailer Sears's centralised business is a large warehousing business along the rail line, and its decentralized business is a variety of retail businesses. Whether centralized, or decentralized, is within an enterprise within the division of labor.

In the cloud computing environment, the business of centralizing business (cloud business) is the base business service provider, such as infrastructure services providers (in the case of IaaS, such as Amazon), platform service providers (in PAAs cases, such as Google, Apple, Alibaba) or software service providers (in SaaS cases, such as Salesforce.com, Ufida), operating a decentralized business (end business), is the application of enterprises (such as Alibaba's owner, Apple ipad developers).

The common denominator of the business is to assume the initial fixed capital investment, and the common denominator of the business is to take the marginal cost into work.

There is an "accident" in economics. Cloud computing is actually in the initial fixed input and marginal investment between the division of labor. Some manufacturers focus on the initial fixed input, engaged in the basic business, and the other part of the manufacturers focus on marginal input, engaged in value-added applications business. And those who are engaged in marginal business can share the resources of the vendors who are engaged in the initial fixed-input business. For example, shop owners share the platform that Alibaba offers free of charge, and ipad developers share the development tools and online stores offered by Apple. Here, the initial fixed input and marginal input unit, is no longer a business, but a commercial ecological community (virtual Enterprise).

This is a phenomenon that the traditional economy does not have. The typical characteristic of the traditional economy is that each enterprise engages in initial fixed capital investment, such as its own procurement equipment, and carries out its own marginal input. When engaged in production, the use of dedicated capital, while the fixed cost and the flow costs, initial fixed inputs and investment in marginal investment. In layman's parlance, it is to build your own factory and produce it yourself.

But in the cloud computing conditions, the enterprise can be someone else to build the factory, their own production. and others cover the factory, you can share. Who wants to produce, share the infrastructure (I, information Infrastructure), "Plant" (P, platform) and "equipment" (S, software), and share resources through the server. Of course, sharing is not white sharing, in the case of Apple, the relationship between the two sides is 37 divided. Provides 30% of server-shared resources, and 70% for application developers. The equivalent of the producer in the leasing of plant and equipment, service providers in the provision of plant and equipment leasing services.

Why is this happening? This comes down to the different characteristics of the means of production. First of all, plant and equipment to be universal. This is difficult for the traditional economy. You can't rent a carpenter's hand drill to dig wells, because the equipment is not common. The entity of the production equipment, which causes its specificity is strong, but the universality is weak. People can rent the means of production, but due to the limitations of universality, it will not be too general. And the means of production of cloud computing is fictitious, the versatility is strong, and the specificity is weak, the ipad can carry on a variety of application development, make the tools are the same. Secondly, cloud computing infrastructure, platform and software, can be reused, many times, its use value will almost not wear. For example, the more shopkeepers on Taobao platform, the higher the value of the platform (equivalent to the more steamed bread more).

Complementary means, the focus of business (basic business, initial fixed input) to share the way to "complement" decentralized, decentralized business (value-added applications, marginal input) to pay rent to the way to "fill" the focus side. Both. Its economic characteristics are scale economy and scope economy, that is, allocating resources in a shared way to allocate resources, the total effect is greater (that is, the scale or variety of one plus one is greater than two, approximate to the cost of additive, cost weak additive).

3, the connection is very important

Focus and scatter, this is the two basic points of cloud computing. They correspond to initial fixed input and marginal input, basic business and value-added service, platform enterprise and application Enterprise respectively. The reason why we cannot separate the two is because once separated, the economic effect of the cooperation between them is no longer.

In economics, the increment of compensation must be connected with both the initial fixed input and the marginal input, but it is impossible to talk about the initial fixed input or the marginal input in isolation. When the industry talks about cloud computing, one does not pay attention to it, it often separates the two basic points. Not to pull cloud computing to the direction of centralized computing, is to the direction of distributed computing, so that people ignore the relationship between the two.

The first kind of fragmentation is a stark split, mentioning only one aspect of cloud computing, without mentioning another.

For example, some people say that cloud computing is a grid technology, which only captures one side and throws the other side. The failure of grid technology lies in the separation of applications and the concentration of business in one direction. Today, grid and the Internet of things can not be isolated development, should be included in the full framework of cloud computing development, so as the application of marginal input, and they form complementary to play their role in the initial fixed input. If only by the government action, the mode of public goods to advance, will commit the former science and technology departments often make a common problem, as a result of separation from the application of market economy and nobody.

The second kind of fragmentation, is a relatively covert fragmentation, cloud computing two aspects are mentioned, but changed the linkage mechanism.

Let's dissect Microsoft's "cloud plus end" model to help understand this fragmentation. Microsoft's proposed "cloud-side", in my view, is actually opposed to cloud computing, but not to mention, but to modify the definition of cloud computing, to negate the common sense of cloud computing. In the past, "Da Jung" means this. Microsoft, in the name of cloud computing, and the point of denying the reality of cloud computing, is focusing on the fragmentation of relationships in software services.

In the case of software services, according to normal understanding (such as Google and IBM understanding), the relationship between cloud and end is software free, service charge, that is SaaS. and Microsoft's "cloud plus end" means that software charges, services are also charged. This has castrated cloud computing from the root. The background is that Google and IBM software are open source software, while Microsoft software is licensed to charge software. Google and IBM advocated the idea of cloud computing, that is, open source software as the initial fixed input, free to everyone, and then through the marginal application services, the cost back. The rationale for service charges is that initial fixed input lowers the cost of each service (no more than one by one initial fixed inputs). Call Microsoft a "correction", it becomes, the software itself through the licensing mechanism to collect a fee, if additional services, the service to collect a fee. Software here, not only to lock the user role, but also charge.

Microsoft also called itself the cloud, the logic is: because I put the service in the cloud, so I am also cloud computing. It is implied that the initial fixed input on software does not reduce the marginal input cost of the application end, and there is no substantial correlation between them. If the software charges, then the software "equipment" on the initial fixed input, no longer to share the way to reduce the total cost of the application side.

This is the same as in traditional economics, where the rules of the game of proprietary capital allow people to pay for both initial fixed costs and marginal costs. Therefore, it does not conform to what we call "cloud computing, which separates the initial fixed input from the subject of marginal input, and joins in complementary ways."

The connection between the cloud and the end is an important issue in cloud computing. The API (application interface) has historically been a bridge between centralized systems and decentralized applications. Cloud computing people do not have the API as an inevitable component of cloud computing, but whether from the mechanism or the practical needs, cloud computing in the cloud and the end of how to achieve the mechanism of the connection, is a problem that can not be ignored. Whether it's in the public cloud, the private cloud, or IaaS, PAAs, and SaaS, the combination of concentrated resources distributed in the cloud and a wide variety of local applications determines the success or failure of cloud computing.

The connection mechanism called the API does not have any relation, the key is to solve the connection problem from the mechanism. The core of the join problem is two aspects, one is the system-oriented and the overall architecture, the other is the local computing and interconnection. Looking back on the history, the problems that middleware has to solve are the same in cloud computing. If the join mechanism can only be oriented to the former, it solves the problem of terminal customization, and solves the problem of standardization and cost if it can only face the latter. On the contrary, history also shows that the connection between the system and the application of the problem solved well, can become passive as active, and even turn the tables. The success of the Apple API, for example, has led to a transformational change. API is not only an interface problem, it is a system engineering, from development tools to business models, involving all aspects of the problem.

To say, the API is actually a large-scale industrial civilization with personalized customization of the agricultural civilization of the borrowing, if the operation is good, the formation of mass customization, the effectiveness of new commercial civilization can play to the extreme, cloud computing will shine.

Cloud computing can be explained in many ways, from an economic point of view, and this is just opinion.

(Responsible editor: admin)