The future of cloud computing

In October 2016, Alibaba Cloud announced a comprehensive reduction in the price of cloud products in China, with a maximum drop of 50%. This is the 17th price cut by Alibaba Cloud in one year. It is everyone worried that this price cut will cause cloud computing prices. During the war, there was an interesting news:

Zhang Jie, a commentator on the China Economic Net, said in the "My Finance" program that this phenomenon reflects deeper problems, and policy makers in China should think more. Cloud computing is calculated by using the rich resources of the client computer. After installing various softwares, the original text processing is performed. The computer's idleness is 70%, the busyness is 30%, and now the basic computer busyness is 70. More than %, these are all used to do cloud computing yet. There are many Chinese computers, the relevant legal thresholds are the least, the cost is low, and the foreign legal restrictions are many, and the cost is definitely high.

Zhang Jie pointed out that in the past, the ability to use the surplus of the client computer to do calculations for others did not allow the customer to fully understand that the customer could not receive the fee and could not share the money. We called it rogue software. Nowadays, it is called cloud computing. In some places in foreign countries, there are legal restrictions, and customers must be allocated money. However, when some companies in China use other people’s computer resources to carry out cloud services, they do not give users any money, which is why the cost is low. And the reason for the price reduction.

——News content comes from the network

Some people who know each other will pass by, and some people will forward their friends circle to express ridicule; people who don’t understand don’t know what’s going on, and believe that there should be quite a lot.

However, looking back, the essence of cloud computing is not computing resources and storage resources. If end users have computing resources or storage resources that can be used, why not use them?

No, no, we look at two concepts first:

Distributed Computing

Distributed computing is a computational method that is relative to centralized computing.

With the development of computing technology, some applications require very large computing power to complete, and if centralized computing is used, it takes a long time. Distributed computing breaks down the application into small parts that are distributed to multiple computers for processing. This can save the overall calculation time and greatly improve the calculation efficiency.

Parallel Computing

Parallel computing is relative to serial computing. It is an algorithm that can execute multiple instructions at a time, with the goal of increasing the speed of calculations and solving large and complex computational problems by expanding the problem solving scale.

The so-called parallel computing can be divided into parallel in time and parallel in space. Parallelism in time refers to pipeline technology, while parallelism in space refers to performing computations concurrently with multiple processors.

As you can see, parallel computing is also a distributed computing, but it is more demanding than distributed computing.

Distributed computing seeks to maximize computing resources, and parallel computing places the highest computational efficiency. Distributed computing that just wants to take advantage of client computing resources in the hands of users already exists.

The SETI@home program runs on the user's personal computer, usually in screen saver mode or in background mode. It utilizes redundant processor resources and does not affect the normal use of the computer by the user. The SETI@home project has been officially launched since May 17, 1999. As of May 2004, nearly 5×10E21 floating-point operations were accumulated, and more than 1.3 billion data units were processed. By the end of 2005, it had attracted 5.43 million users, and their computers had accumulated 2.43 million years of work. , analyzed a large amount of backlog data.

SETI@home is a scientific experimental program for the search for extraterrestrial civilization (SETI) using a globally connected computer. It was founded by the famous American university of UC Berkeley. The center platform is set up at the Space Sciences Laboratory (SSL). ). Volunteers can join the project by running a free program to download and analyze the data from the radio telescope.

For more information, you can check the official website of the project http://setiathome.berkeley.edu/ 

You see that only users can voluntarily load a certain amount of data through a small program, and some simple distributed computing can be done, and distributed computing here occurs earlier than large-scale centralized parallel cloud computing.

In addition to looking for aliens and gravitational waves, there are more down-to-earth distributed computing projects. Solve more complex mathematical problems, such as: GIMPS (find the largest Maison prime). Research to find the most secure cryptosystem, such as: RC-72 (password cracking). Biopathological studies, for example: Folding@home (research on protein folding, misunderstanding, polymerization and related diseases). Drug research for a variety of diseases, such as United Devices (search for effective drugs against cancer). Study the physiological principles of AIDS and related drugs, FightAIDS@home and so on.

The development of this kind of distributed computing technology has gradually formed a standardized platform service, such as BOINC, World Community Grid (WCG) and other distributed computing platforms, which can let more people feel the benefits of distributed computing.

BOINC is an acronym for Berkeley Open Infrastructure for Network Computing, the Berkeley Open Network Computing Platform, an open middleware system for volunteer computing and grid computing. Originally developed to support the SETI@home project, BOINC has since become the most mainstream distributed computing platform for a wide range of projects in the disciplines of mathematics, physics, chemistry, life sciences, and earth sciences. BOINC was developed to help researchers more easily access computing resources from volunteers around the world. The future of BOINC is very impressive and it is possible to develop into an industry standard. With the BOINC platform, the development and promotion of distributed computing becomes easier and easier. The unified interface, unified way will greatly facilitate the new users of distributed computing, without having to study the participation methods and integration algorithms of each different project. BOINC has matured and there are currently hundreds of projects that have successfully run on the BOINC platform.

World Community Grid (WCG) is a distributed computing platform that allows you to participate in the research of multiple projects. Research organizations can submit their own research topics on the project application page to get free computing resources from World Community Grid. World Community Grid uses the computer's idle computing power to do some engineering that benefits all humans. With screen savers, grid technology will be easier to use, safer and more free. When you need to use your computer, the grid computing software will automatically stop running until your computer is idle again. It is a global philanthropy business (initiated by IBM and some research, charitable, educational organizations) that uses unused computing power in personal and business computers for philanthropy.

This type of distributed computing basically has some common features, such as easy to slice, issue, and recycle tasks. The final calculation time is not so high. The real-time and correlation between calculations is not strong. Finally, the program through the client is used. (such as screensavers) to achieve distributed computing on a personal computer, rather than a real strong correlation, real-time parallel computing. At present, the above distributed computing is of a public nature, that is to say, it does not take the user's computing resources to make a profit.

Alibaba Cloud 2017 National Cloud Computing Activity

However, the current cloud computing, including AWS, Alibaba Cloud, etc., provides parallel cloud computing, which is not only about distributed, but also the efficiency of operation, the high availability of the operating environment, and so on.

Is there any possibility for this kind of distributed computing to develop into the current parallel cloud computing? With the advancement of technology, it is theoretically possible to solve such a problem, but whether the user is willing, whether the technology to solve the problem is mature, and whether the cost is cost-effective.

If the technology advances, and the user can use the computing and storage resources in the hands of the user to achieve true parallel computing, the world's computing resources can be effectively utilized to achieve universal cloud computing. Of course, the basis of this model still requires the end user's permission. If the end user is unknowingly used, it is often rogue software, or a virus, such as a zombie computer used for DoS attacks. ), referred to as "zombie", some people call it "broiler chicken". After the computer connected to the Internet is infected by a virus, it is controlled by a hacker. At any time, it can launch a denial of service (DoS) attack or send spam according to the instructions of the hacker. .

Let us rethink a question, why do today's cloud computing vendors provide scalable public cloud computing services, and even we are still envisioning the use of all available computing and storage resources to achieve true universal cloud computing?

There is only one reason. I am worried that local computing resources are not enough. Even local private clouds have limited resources. Public cloud computing resources and even universal cloud computing resources are extremely rich and flexible.

Even if we now create a computer with enough computing power and storage capacity, I feel that as long as his computing power and storage capacity are strong enough, public cloud computing is not needed. This is a very beautiful vision, but don't forget one thing, the demand for computing and storage is also infinitely expanding. On one day, when a computer or even a local cluster (private cloud) can't meet the computing and storage needs, it will eventually return to the path of public computing for elastic computing!

In the era of cloud computing and big data, there will be more and more data, especially when the data explodes at an exponential level, the calculation will become more and more complicated.

Da Liu’s god-level civilization, described in Shi Yun, wants to save all possible verses written in Chinese, and almost has to spend the entire sun and devour all the resources of the empire. For another example, in the future, in order to be able to copy or resurrect humans, we will definitely want to store the arrangement of each atom in the human body! Just as people's desires are endless, we have the same demand for computing and storage. I never believe that limited resources can solve infinite demand, and public cloud computing can solve elastic expansion.

Public cloud computing is bound to exist for a long time, and it is the main mode of the future. The final two points are summarized:

1. Reversely sending data to the client terminal. It is not impossible to use the remaining computing power of the client to complete larger distributed computing. It has been used in many occasions. With the advancement of technology, this mode will become more and more Widely used for some specialized computing, even parallel cloud computing, to achieve true national cloud computing. It is currently solving the problem of increasingly inadequate computing and storage resources.

2. Even if there is a possibility of a private cloud, the connection to the public cloud is also essential, that is, at least the "hybrid cloud" model is a predictable long-term existence. Even with the rapid development of computing speed and storage technology, don't expect private clouds to carry all computing needs, because computing and storage requirements are also growing rapidly. Public cloud computing is indispensable and is the future trend.