Shulou(Shulou.com)06/01 Report--

Cloud data center architecture was last updated on the blog in May last year, and it has been eight months now. No excuses are lazy. I have always thought that only the knowledge that can be shared can really belong to me, and what I share is my own. Recently, in the Huawei data center certification course, I have also read the "Bank Information system Architecture" and "Commercial Bank Private Cloud Design and implementation" series of books. I have some ideas for the modern cloud data center architecture. everyone is constantly exploring in this world, and the process of learning is constantly refreshing their cognitive domain, which may suddenly become clear one day. I also hope that this sharing can give some inspiration to the brothers who are interested in learning. Data center is divided into traditional data center IDC and modern data center CDC (cloud data center). Now the vast majority of data centers have been transformed from IDC to CDC. These changes are not much reflected in the underlying infrastructure. The difference is that CDC manages the computing, network, storage and other resources of the underlying layer with software, and then integrates them into the same cloud resource pool. To put it bluntly, it brings all kinds of capital sources together to form a cloud. (at present, the system software that manages all kinds of CDC resources is openstack. Because the code is open source, various manufacturers update and optimize cloud management software based on openstack, such as Huawei's "Fusion Sphere OpenStack". Traditional data centers have encountered many problems, such as the decline in resource utilization, the fact that there are many types of equipment that are difficult to manage in different places, and the slow speed of business launch. On the other hand, the cloud data center introduces the concepts of "resource pooling" and "automation". All kinds of resources are pooled first, and then resource automation is done on a secondary basis. Automation essentially improves the launch speed of new business, so that the business can quickly respond to this change when users' needs change. At the same time, it is also very convenient to expand the capacity of data center resources in the later stage. In CDC, all kinds of lower-level resource virtualization software (such as computing virtualization, storage virtualization, network virtualization) can be driven by the cloud platform, and finally the underlying hardware can be driven by all kinds of virtualization software. To put it bluntly, C DC deploys the software layer on the basis of the underlying hardware and relies on software-driven hardware to achieve cloud data center, that is, SDDC, which is proposed by many manufacturers, that is, software-driven data center. Now there is a new concept called SD (software driven) CD (distributed cloud) DC (data center), there are data center disaster recovery, 2 places 3 centers, multi-activity and so on, so the concept of distributed data center is also put forward. One of the prominent features of cloud data centers is virtualization, and the utilization of device resources will be greatly improved after virtualization. Secondly, cloud data centers should have cloud OS, and cloud OS does two things (resource pooling and business deployment automation). The technical architecture of cloud data center is divided into five layers: 1) hardware infrastructure layer: such as server, network, storage and other hardware facilities. 2) Virtualization layer: deploy a virtualization layer on top of hardware resources (virtualization can slice computing, storage, network and other functions. To make multiple virtual machines share is to slice resources. For example, a CPU virtualizes N CPU, overallocates limited memory, turns virtualization into virtualized slice resources, and integrates these resources again. The virtualization layer is hardware dependent. 3) Cloud resource layer: integrate three types of resources through Cloud OS, and Cloud OS will further pool resources, that is, multiple (such as computing, storage, network) and multiple (such as Huawei virtualization resources, VMware virtualization resources) virtual resource pools are further integrated into a cloud resource pool under secondary unified management. A friendly man-machine interface is added on top of the resource pool, which allows users to operate graphically. Let users see cloud products such as cloud disk, cloud host, LB, router and so on. The cloud products that users can choose are similar to those selected by JD.com. If you choose a CVM, if you have a complex network, you can apply for a router and LB to share the load. 4) Cloud service layer: that is, a layer called cloud service layer that provides cloud products such as CVM and cloud disk for users to use, which provides service products. The service layer connects with the cloud resource layer, and the cloud resource layer calls the virtualization layer. The service layer provides a web page to the user, and the user uses the web page to use the resources in the resource pool. 5) tenant (user): that is, the user pays according to the use of time, so the word tenant comes from. After bringing the sliced resources into the cloud resource pool, it will be further abstracted, shielding all the differentiated characteristics of the vendors. What you see in the cloud resource pool are general-purpose and de-differentiated resources, and the ultimate goal of pooling resources is for users to use. The user is on the northernmost side and the hardware is on the southernmost side. The above is my understanding and understanding of the cloud data center architecture. The blog content will be updated irregularly in the later stage. In the next issue, we will talk about hot VXLAN. As soon as I looked up, the sky was already dark, and the night sky in Shenzhen was as beautiful as her.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.