No.1 What is cloud computing?
Cloud computing is a technology that manages real physical devices through virtual software to meet the flexible requirements for hardware facilities. For example, Ding Dian, the siege lion, wants to calculate 50 million data samples, but his laptop can't handle it, and he doesn't have money to buy more servers to provide What about CUP, memory and hard disk? If anyone can rent their hardware resources once, the siege lions can pay a small fee after calculating, which saves money and meets the demand. Like car rental companies, is there any company that can provide such rental computing resources?
Of course, there are cloud vendors. Cloud manufacturers have bought many servers and set up data centers to provide computing service leasing, but how to provide leasing services? The Siege Shidingdian needs to rent servers in Hangzhou, and the cloud manufacturer's data center is in Beijing, so you can't Please move the server to Hangzhou or Ma Nong A to Beijing. In order to solve this problem, related technologies have emerged to provide security, namely Virtualization and Cloudization 。
No.2 What exactly does cloud computing manage?
It uses virtual technology to manage physical hardware Computing resources (cup, memory), storage resources (hard disk) and network resources (broadband) 。 Use virtual software to simulate a computer and apply for certain resources for use.
In fact, cloud users use the servers in the background of cloud manufacturers, and only allocate some resources from the servers to you through virtual software. This process is Virtualization.
What are the virtualization software?
Closed source includes VMware, and open source includes virtualization software, one is called Xen, the other is called KVM.
No.3 How to allocate cloud resources
Tens of thousands of servers in the background respond to cloud users several times as many as the server. How to quickly and effectively allocate user applications is automatically completed through the algorithm development of a scheduler. The process of scheduling resource allocation and computer configuration is called pooling or cloudization,
Therefore, the scheduling software also becomes cloud software.
Tens of thousands of servers in the background respond to cloud users several times as many as the server. How to quickly and effectively allocate user applications is automatically completed through the algorithm development of a scheduler. The process of scheduling resource allocation and computer configuration is called Pooling or Cloudization ，
Therefore, the scheduling software also becomes Cloud software 。
No.4 Cloud classification
Cloud includes private cloud and public cloud. Private cloud means that enterprises buy hardware equipment to build data centers, purchase virtual software and cloud software to install on their own servers for use by their internal employees.
Public cloud is to rent its own computing resources to ordinary users in the outside world and collect fees.
Cloud computing originally wanted to achieve flexibility at the resource level. This service is at the infrastructure infrastructure level, known as IaaS cloud platform or IaaS cloud computing Infrastructure As A Service 。 Generally speaking, it is a service to rent equipment to make money.
However, the realization of resource flexibility does not really liberate you. In short, there is no software installed on your computer and you need to install it yourself. To solve this problem, cloud vendors have Provided software application services ， PaaS（Platform As A Service）， Provide users with general or customized software. Cloud manufacturers develop tools for automatically deploying software to provide application services, including Puppet, Chef, Ansible, Cloud Foundation, and the latest container technology Docker can do this better.
Container, simply understood as a container, features packaging and standards. Function is to encapsulate various components to form a common tool. At the beginning, the loaders (container developers) put various goods in containers, and the transport workers (container users) transferred the goods In the process, the goods are transferred in containers, which reduces the workload and improves the efficiency.
Upgrade of cloud computing under big data
Cloud computing has experienced the evolution of Iaas and Paas in the early stage, but in the era of big data, this does not meet the demand at all. Iaas provides computing resources and Pass provides application platforms. Now there are new requirements. As a small business, I want to identify the yellow information on the website, and What should I do if my own data cannot meet the model training requirements? Cloud manufacturers seize the business opportunity again. Cloud manufacturers have rich data, use them to train models with high accuracy, and then sell the models to users. Just use my model to analyze your data directly It will be very convenient to carry out model training again. So there is a third version of cloud computing, Saas (Software As A Service) 。 AI programs (models) enter cloud computing as SaaS platforms.
So far, cloud computing, big data and AI have formed a complete ecosystem.
