Cloud computing is the development of distributed computing, parallel computing and grid computing, or the commercial realization of these computer science concepts.
cloud computing
The basic principle of cloud computing is that by distributing computing on a large number of distributed computers instead of local computers or remote servers, the operation of enterprise data centers will be more similar to the Internet. This enables enterprises to switch resources to required applications and access computers and storage systems as needed.
This is a revolutionary measure, for example, it is like changing from the old single generator mode to the centralized power supply mode of power plants. It means that computing power can also be circulated as a commodity like gas and water and electricity, which is convenient to obtain and low in cost. The biggest difference is spread through the Internet.
The blueprint of cloud computing has come to the fore: in the future, everything we need, even such tasks as supercomputing, can be realized through network services, only a notebook or a mobile phone is needed. From this perspective, the end user is the real master of cloud computing.
The application of cloud computing includes the idea of uniting strength for each member.
Judging whether it is the three criteria of cloud computing.
What is cloud computing?
What exactly is cloud computing? This is a problem that everyone is more concerned about. Now we find that there are many different views. What is a cloud and what is not a cloud is puzzling. Some people say that the public cloud is a cloud and the private cloud is not a cloud; Others say that supporting virtualization is called cloud, and not supporting virtualization is not called cloud, but Google does not support virtualization, and we all think that Google is cloud; Others say that 1000 servers are clouds, as if 999 servers are not clouds. Now there are some high-performance computing centers, nothing has changed, but the name has been changed to cloud computing center. Why do so many people use the banner of cloud computing? Either earn eyeballs or circle money, each with its own purpose.
As the public, we need to distinguish which clouds are true and which are false. To this end, Professor Liu Peng put forward three reference standards to judge whether it is cloud computing, for your reference [4]:
Users need resources from the network rather than the client.
This is the basic idea of cloud computing, that is, providing users with computing power, storage space, software functions and information services through the network.
The service capability can be extended in minutes or seconds.
If the service capacity of resource nodes is not enough, but the network traffic comes up, the platform needs to automatically and dynamically increase the number of service nodes from 100 nodes to 150 nodes within a few minutes. If it can be called cloud computing, we need enough resources to cope with the peak of network traffic, even if hot spots like "Jia" suddenly appear. After a period of time, the traffic came down, and the number of service nodes decreased with the decrease of traffic. At present, some traditional IDC claims to provide scalability, but it takes many hours to provide it to users. The problem is that the network traffic is unpredictable and it is impossible to wait that long.
Compared with the traditional model, it has more than 5 times the cost-effective advantage.
After reading the above item, some people are thinking, it doesn't matter, one more machine can cope with large traffic. But this is not the idea of cloud computing. We also have cost-effective indicators. Cloud computing is an epoch-making technology, because it puts a large number of cheap computers into a resource pool, uses software fault tolerance to reduce hardware costs, saves electricity costs by deploying cloud computing facilities in cold and electricity-rich areas, and improves resource utilization through large-scale use. Representative cloud computing platform providers abroad have achieved an amazing cost performance ratio of 10-40 times. Due to domestic technology, scale, unified electricity price and other issues, it is difficult to achieve the same cost performance for the time being, so we temporarily set this index as 5 times. The cloud computing platform of China Mobile Research Institute with 256 nodes has improved its cost performance by 5 to 7 times. With the increase of scale and utilization, there is still room for improvement in cost performance.
example
These three standards are equivalent to three filters, which can filter out things that do not belong to cloud computing. Let's try it-see if the following are cloud computing:
PC system: It can't pass the first level, because all the resources users need are in the client, so it doesn't belong to the category of cloud computing;
IPhone application software: if downloaded to the iPhone, it can run independently, and it is only a communication relationship with the outside world and will not pass the first level; If it runs on a network platform, the computing and data resources come from the network, and the iPhone is just an access interface, it may enter the category of cloud computing;
Website: Pass the first level. However, if the scalability is still passed, the comparison between performance and price level is very difficult;
Netbook: Can pass the first level, but only the terminal. If it is only used for surfing the Internet and interacting with cloud computing facilities, it can only be classified as traditional cloud computing.
Advertising alliance: Sorrow is the second layer. Organizing advertisements from different websites together constitutes advertising alliance. Publish an advertisement that can appear on many websites. However, due to the low click-through rate of advertisements by the public, advertising alliance does not need much expansion, and it is not easy to do it-because the participants are many small websites;
Supercomputing center: if the scale is large enough, it can barely pass the second level, but the cost performance is not good, but it can't pass the third level. Supercomputers are very expensive and have no advantage in cost performance. We don't think it is a cloud;
Online office: It is easy to meet the requirements of cloud computing after large-scale operation. For example, the Parkway Office you saw today, and Google's DOCS.
Online CRM: It is easy to meet the requirements of cloud computing after large-scale operation. Tens of dollars to provide users with a month's use, far superior to the traditional model we use, especially in the cloud;
IaaS: Like Amazon's rental service, so this kind of cost performance is also very good. Renting a virtual machine 1 hour is just 0. 1 dollar, which is also a cloud.
Paas: For example, Microsoft's Azure can provide 25GB of storage space for free, which is definitely very cost-effective, and it is also a cloud.
Cloud security: It is easy to meet the requirements of cloud computing after large-scale operation. For example, 360 security guards provide users with completely free services, which is possible because of its high cost performance.