Distributed computing is a computing method, which is opposite to centralized computing.
With the development of computing technology, some applications require huge computing power to complete. If centralized computing is adopted, it will take a long time to complete.
Distributed computing decomposes the application into many small parts and distributes them to multiple computers for processing.This can save the overall calculation time and greatly improve the calculation efficiency.
Study how to divide a problem that requires huge computing power into many small parts, then allocate these parts to many computers for processing, and finally integrate these calculation results to get the final result.
The recent distributed computing project has been used to use the idle computing power of the computers of thousands of volunteers around the worldInternet, can analyze the electrical signals from outer space, find hidden black holes, and explore possible extraterrestrialsSmart life；You can find Mason prime numbers with more than 10 million digits;We can also find more effective drugs against HIV.These projects are very large and require an astonishing amount of calculation. It is impossible to complete the calculation only by a single computer or individual in an acceptable time.
Definition of Chinese Academy of Sciences
Two or more software can share information with each other. These software can run on the same computer or on multiple computers connected through the network.Compared with other algorithms, distributed computing has the following advantages:
1. Rare resources can be shared.
2. Distributed computing can balance computing load on multiple computers.
3. The program can be placed on the computer most suitable for running it.
Among them, sharing rare resources and balancing load is one of the core ideas of computer distributed computing.
Grid computing
Grid computingIt is a kind of distributed computing.If a work is distributed, it must involve not only one computer, but also onecomputer networkThis "ant moving mountain" method will have a strongdata processingAbility.
The essence of grid computing is to combine and share resources and ensure system security.
working principle
Distributed computing usesinternetOf computer ona central processorIt is a kind of computing science to solve large-scale computing problems with idle processing capacity of.Next, let's see how it works:
First, we need to find a problem that requires huge computing power to solve.Such problems are generally interdisciplinary, challenging and urgent scientific research topics for human beings.The more famous ones are:
one
Solve more complex mathematical problems, such as:GIMPS(Look for the largestMersenne prime ）。
two
Research to find the safestCryptosystem, for example:RC-72(Password cracking).
three
Biopathological studies, such as:Folding@home(Research on protein folding, misunderstanding, polymerization and related diseases caused by them).
four
Drug research for various diseases, such as:United Devices(Looking for effective drugs against cancer).
five
Signal processing, such as:SETI@Home(Looking for extraterrestrial civilization at home).
Previously, these problems should be solved by supercomputers.However, the cost and maintenance of supercomputers are very expensive, which is not affordable for an ordinary scientific research organization.With the development of science, a cheap, efficient and easy to maintain computing method - distributed computing!
With the popularity of computers,personal computerStart to enter thousands of households.With it comes the use of computers.More and more computers are idle, and the potential of the CPU is far from being fully utilized even when the computer is turned on.We can imagine that a home computer spends most of its time "waiting".Even when users actually use their computers, the processors still consume silently and wait countless times (waiting for input, but actually doing nothing).With the advent of the Internet, the connection calls all those who have limited computing resourcescomputer systemIt has become a reality.
Then, some very complex problems that are suitable for dividing into a large number of smaller pieces of computing are proposed, and then a research organization develops computing servers andclient。The server is responsible for dividing the computing problem into many small computing parts, and then distributing these parts to many computers connected to the network for computingparallel processing Finally, these calculation results are combined to get the final result.
Of course, this seems to be primitive and difficult, but with the increasing number of participants and computers involved in computing, computing plans become very fast and are proved to be feasible by practice.At present, the processing capacity of some large distributed computing projects can reach or even surpass the fastest supercomputer in the world.
You can also choose to participate in some projects to donate the CPU core processing time. You will find that the CPU core processing time you provide will appear in the project contribution statistics.You can compete with other participants to rank your contribution time. You can also join an existing computing group or form a computing group yourself.This approach is conducive to mobilizing the enthusiasm of participants.
With the growing number of private teams, many large organizations (such as companies, schools and various websites) have also started to form their own teams.At the same time, a large number of communities with the theme of distributed computing technology and project discussion have also been formed. Most of these communities translate and produce tutorials for the use of distributed computing projects and publish relevant technical articles, and provide necessary technical support.
So who might join these projects?Of course, anyone can!If you have joined a project and have considered joining the calculation team, you willChina Distributed Computing CenterAnd find your home in the forum.Anyone can join any distributed computing group set up by us.We hope you will find fun in China's distributed terminals and forums.
To participate in distributed computing - the most meaningful choice that can give full play to the use value of your personal computer - you only need to download the relevant program, and then this program will run on the computer with the lowest priority, which has little impact on normal use of the computer at ordinary times.If you want to use your spare time to do something useful, why hesitate?Take action at once. Your trivial efforts may help you become a human beingHistory of ScienceThere is a big sum left on it!
BOINC distributed computing platform
BOINC is short for Berkeley Open Infrastructure for Network Computing, namely Berkeley Open Network Computing Platform.
BOINC is a distributed computing platform that can be shared by different distributed computing.Different distributed computing projects can directly use BOINC's public upload and download system, statistical system, etc. This can not only play a role in the coordination between various distributed computing, but also make the management and use of distributed computing more convenient.
BOINC has its ownIntegral systemBecause the projects that can be run on BOINC are very different, for example, the Workunit (WU) of Project A needs 3 hours to complete in a machine, while the task package of Project B needs 30 hours to complete in this machine. Obviously, it is not feasible to use the number of WUs to measure the work volume;Similarly, there are differences in machine performance, and it is not feasible to use CPU time to measure workload.The integration system can only get the calculation amount actually completed by the user through a certain algorithm. This article explains the calculation method of integration in BOINC.
Distributed Computing in China
according toChina Internet Information CenterAccording to the statistics of (CNNIC), the number of Chinese netizens has increased to about 12% of the world, and is still growing rapidly.Chinese netizens mentioned here include mainlandHong Kong, Macao, Taiwan and overseas Chinese.
Compared with the rapid development of the Internet in China, distributed computing has developed slowly in China.The number of Internet users does not objectively reflect the level of informatization in a country. The number or proportion of Internet users participating in distributed computing can show the level of scientific popularization in this country.In this regard, European and American countries are in the lead.stayNorthern EuropeAlmost half of the computers have participated in distributed computing projects, which is quite surprising.In China andIndiaWith many latest technologies, it has also performed well in the process of network popularization, but it is still weak in distributed computing.Let's look at some examples:
SETI@homeIt is the largest distributed computing project in the world and the most involved distributed computing project in China.It searches for signs of life outside the Earth by analyzing the signals received by radio telescopes using online computer download programs.
Although China has made progress in the field of distributed computing: its international ranking has risen from 29 to 24, the amount of computing completed by Chinese users is only one tenth that of Japanese users, and the number of Japanese Internet users is less than that of China.In order to make further breakthroughs, it is undoubtedly very important to improve the quality of Internet users and the degree of scientific popularization.India and someMiddle EastCountries also have similar problems. Central European countries do a better job. Many CIS countries have completed more data calculations than the Russian Federationthe sum。
Folding@homeIt is a distributed computing project that studies protein folding, misunderstanding, polymerization and related diseases caused by it.It uses networked computing and a large amount of distributed computing power to simulate the process of protein folding, and guide us to a series of studies on diseases caused by folding.
The number of Chinese participants in this project is increasing. By the end of March 2011, the number of users of the 3213 team in China had reached 3025, and the number of active users was about 190.
Participating in the Climateprediction project is to calculate the latest climate prediction model through computers in homes, schools and offices.The results of these calculations will form the world's largest weather prediction model.The climate has changed, and our action on it has become an important global topic.This will affect human agricultural production, water resources, ecosystems, energy demand, insurance costs and many other aspects closely related to human beings.There is solid scientific evidence that the earth may become warmer in the next few centuries, but we cannot estimate how big the change will be.If you participate, it will help with the climate of the 21st centuryScientific prediction。
China already has many computers, many of which are extremely advanced in performance.Most of them just type and play slides.This is a waste of resources.
From another perspective, we can easily find thatdeveloped countryanddeveloping countryGap.We call this phenomenon the digital divide.Another phenomenon is alsowith deep hatred and resentmentAll distributed computing projects are developed by developed countries, such as the United StatesGermany, Britain, Japan and so on, which also deepened the scientific gap.Stanford UniversityProfessor Gorham Richard Cherman of the Department of Chemistry said that distributed computing would accelerate the scientific process of the entire human race.Scientists can accomplish computing tasks that were never expected to be completed before, or that will take decades and hundreds of years to complete.This is indeed true, but this calculation has contributed to the scientific monopoly of developed countries to a certain extent.
Relevant departments in China have also begun to realize the importance of distributed computing. Some university professors and scientists have also begun to study distributed computing science, such as the Chinese Academy of SciencesCAS@HOMEandTsinghua University"Clean Water Plan".
Meaning pattern
Announce
edit
Distributed Computing and Human
Due to various subjects of modern human beings, they cover a wide range of subjects and are classified in detail.Nowadays, every subject seems to need a lot of calculation.Astronomical research organizations need computers to analyze space pulse and star position movement;Biologists need computers to simulate the process of protein folding;Pharmacologists want to develop and overcomeAIDS(AIDS) orSars(SARS) drugs;Mathematicians want to calculate the largestPrime numberandPiMore accurate values of;Economists use computers to analyze and calculate the development direction of an enterprise/city/country under tens of thousands of factorsmacro-control。It can be seen that the science of the future of mankind is always inseparable from computing.Distributed Computing, with its unique advantages - cheap and efficient, has attracted more and more social attention.
Distributed computing pattern
At present, there are about 100 kinds of distributed computing in the world, most of which are independent of each other, managed and used their own software independently.This kind of distributed computing pattern is not conducive to the needs of development.such as, somebiologyResearch institutions need to use the computers of volunteers around the world to simulate the process of protein folding. That biological research institution has no professionals in distributed computing, but there is no company in the society that can provide suchservice, they have to spend a lot of energy on developing distributed computingThe server, client.In this way, the time that could be used to study biology was spent elsewhere.The biological research institution just mentioned isStanford UniversityPANDE team.
BOINC unified overall situation
In order to change this chaotic separatism,University of California at Berkeley (UC Berkeley) first proposed the idea of establishing BOINC.The full Chinese name of BOINC isBerkeley Open Network Computing Platform(Berkeley Open Infrastructure for Network Computing).And uniformly allocate computer resources (for example, if you are interested in researching AIDS drugs and exploring extraterrestrial civilization, you can select two operations at the same time and set priorities).Unified management of the statistical scoring system (no matter which project you are working for, as long as you devote a long CPU time, the score will be high).With such unified management, it really provides convenience for scientific research institutions like PANDE Group!
BOINC has matured, and several projects have successfully run on the BOINC platform, such asSETI@home，LHC@homeEtc.
Security
Announce
edit
For the user side, before joining any project, you must ensure that you can trust the developer of the project, which mainly involves two aspects:
Private data on your computer
The computing program you downloaded from the project party runs on the local computer and can access the network. Therefore, only a reliable project party can ensure that the private data on your computer will not be taken away or modified maliciously.
Life span of personal computer
Although the computing program of distributed computing generally runs at the lowest priority, which will not affect your daily use, it will still exert some pressure on various parts of the computer when the computing program runs at full load. For more information, please check the impact of distributed computing on computer hardware and software.
For the project side, the volunteers participating in distributed computing are not the project side's own personnel after all, and not all can be trusted. Therefore, a certain redundant computer system must be introduced to prevent calculation errors, malicious cheating, etc.
