Cloud computing in the past and this life

Why "cloud"?

"Everything is about demand and using existing processing power and resources." "

Moore's Law, published by Moore, founder of Intel Partnership, in 1965, predicts that the number of transistors used in a processor will double every two years. The past 40 years have proved the correctness of this law, and the law is likely to remain true at least 2020 years ago.

If you compare today's typical handheld devices (such as iPad2), its A9 dual-core processor is twice times the Intel Pentium III processor's computing power, which was commonly used in the Wintel server in 2000. Today, the new Wintel server is about 40 times times more capable. So what does that mean?

Back to 2000, the case of running a set of Windows server operating systems on each server is typical, and there is one or more applications and services on top of the server. As a result, a company may eventually need multiple servers to provide a full range of IT services, such as AD services, Dhcp/dns services, Microsoft Exchange Mail Services, Microsoft SQL database services, file archiving services, and so on. However, over time and with the rapid increase in processor power, the software industry has not developed at the same speed to give full play to processor power: there is no significant innovative application service software that requires 10 times-or 20 times-fold processor computing power. Therefore, the use of powerful hardware to run a few applications is becoming increasingly wasteful, but running multiple applications on a single operating system can create conflicts. This is how virtualization is born.

Virtualization is simply the ability to have multiple servers on one hardware. By running virtual machine management software on hardware, virtual machines (VMS) provide virtual, independent space for each application and service. In addition, a crashed virtual machine does not cause other virtual machines to fail through the chain reaction. Top virtual platform providers are: Core (VMWare), Jie (Citrix) and Microsoft (Microsoft). Although there are other virtual machine platforms, they are designed for specific operating systems, such as Linux operating systems only, rather than virtual hardware that can run any operating system. The benefits of virtualization are the ability to leverage existing processor resources more fully, reducing the number of hardware enclosures and the consequent reduced footprint, power, and heat.

However, virtualization quickly encounters a design flaw in multiple (virtual machine) to one (hardware). Multiple virtual machines share a single piece of hardware and often require tedious steps to transfer a virtual machine from one piece of hardware to another. There is an argument that the traditional "one-to-one" is like living in a big house, and that the virtualization of "Many-to-many" is just like living in an apartment. While living space is more effectively used, you must manage the many tenants living in the apartment to ensure they don't fight each other. Obviously, the laws of natural evolution focus on "Many-to-many"-cloud computing.

In simple terms, multiple virtual machines share the resources of multiple hardware, and these hardware links form a "cloud". There are many different interpretations of the term "cloud computing", such as:

"Internet Computing"-which also includes the private Cloud, because it also uses Internet protocols and technologies.

"Utility computing"--each service is charged according to utility, i.e. the actual resource used. There is also the use of a service-oriented architecture (SOA), in which customers pay for services rather than for hardware.

"Grid Computing"--cloud technology is a form of distributing operational processing to available resources.