Cloud Computing Layered Architecture

The Internet of Things and the integration of three networks have greatly promoted the development of networks. The use of 3G terminals has enabled wireless and real-time access to the network, facilitating the interconnection of voice, data, images, and other information. This will rapidly expand both the customer base and the total amount of information, inevitably presenting a geometric explosion of development opportunities for the network.

Cloud Computing

HierarchyArchitectureWang Hongliang

Among them, cloud infrastructure as a service (IaaS), cloud platform as a service (PaaS), and cloud software as a service (SaaS) are service models of cloud computing.

Due to data centers often purchasing various devices from different manufacturers simultaneously, and the coexistence of new and old equipment, a significant characteristic of this field is the presence of a large amount of additional equipment resources.

Through basic software and hardware monitoring, distributed file systems, and virtual computing, the cloud operating system is responsible for the management and monitoring of basic software and hardware resources in cloud computing data centers. It ensures data and information security through a security management center that enables multi-user sharing of resources, and effectively achieves green and low-carbon operations of basic resources through an energy management center. The business and resource scheduling center realizes part of the content of the cloud platform as a service layer.

Service Models of Cloud Computing

　　
Cloud Infrastructure as a Service (IaaS, Cloud Infrastructure as a Service)

Cloud infrastructure as a service allows system vendors to provide users with measurable computing, storage, network, and standalone operating system resources of the same granularity, on which users can deploy or run various software, including client operating systems and application businesses.

Cloud Platform as a Service (PaaS, Cloud Platform as a Service)

Cloud platform as a service refers to the cloud computing platform vendor providing the development environment and operating environment for business software as a service to users via the internet. Cloud platform as a service needs to be built on top of cloud infrastructure. Users can create their business applications in the development environment provided by the cloud platform vendor and can operate their businesses directly in the operating environment of the cloud platform.

Cloud Software as a Service (SaaS, Cloud Software as a Service)

Software as a service refers to a software application model where operators provide software services to users via the internet. Traditional SaaS and cloud SaaS are similar in customer experience, such as Sina Mail and Gmail, where customer experiences are comparable. Traditional SaaS is built directly on hardware devices, unable to achieve multi-tenant sharing of backend resources or dynamic flow of resources, and does not belong to the scale of cloud computing. Cloud SaaS requires these software businesses to operate on the cloud platform service layer or be built on the cloud infrastructure layer. The advantages of cloud SaaS are reflected in the dynamic scalability and flow of backend resources, which is a significant advantage that traditional SaaS does not possess.
　　

Cloud Computing Deployment Models

Private Cloud: A private cloud refers to a cloud platform used by a single organization internally. Private clouds are the most common deployment model when enterprises build cloud platforms internally.

Community Cloud: A community cloud refers to a cloud platform shared by several related organizations.

Public Cloud: A public cloud is a cloud platform that can be used by the public or a large industrial group, generally owned by an organization that specializes in selling cloud services.

Hybrid Cloud: A hybrid cloud is a combination of two parallel multi-clouds (private cloud, community cloud, public cloud) that retains their independent cloud entities, which are bound together through standard or private data and business migration technologies.
　　

Cloud Computing Hardware Infrastructure

The Internet of Things, the integration of three networks, has greatly promoted the development of networks. Through 3G terminals, wireless and real-time access to the network has been achieved, bringing interconnectivity of voice, data, images, and other information, which will rapidly expand the customer base and total information volume, inevitably leading to a geometric explosion in the potential for network development. The development of networks requires data centers to have stronger data processing capabilities and also demands green and energy-efficient operations for increasingly large basic resources.

Cloud computing is appropriately aligned with the future higher-level development trends of networks, focusing on improving network data processing and storage capabilities, and committed to the low-carbon and efficient use of basic resources. Specifically, it should focus on enhancing the data processing capabilities of cloud computing data centers in the fields of high-end servers, high-density servers, massive storage devices, and high-performance computing devices.

High-End Servers

Cloud computing faces numerous concurrent businesses, massive information storage, involves massive data retrieval, analysis, processing, data concurrent processing, and data diversity, requiring strong computing power to support.

Compared to overall scale, high-end servers have advantages in shared memory, system limits, stability, high performance, and high security, making them more suitable for core applications in commercial computing and future data centers.

High-end servers undertake high-load tasks such as cloud computing platform management, resource scheduling, and job scheduling, making them an inevitable choice for core cloud computing businesses and the core devices of the entire cloud computing system. For example, in cloud storage systems, although storage capacity can use inexpensive servers, the main servers for managing namespace, access control, data retrieval, and mapping relationships from files to data blocks, especially in distributed file systems facing massive data, often require high-end servers to ensure the efficiency, availability, and robustness of the distributed file system.

High-Density Servers

For the numerous ordinary applications undertaken by cloud computing platforms, a large number of standard servers are needed to meet green and low-carbon application demands. Traditional rack servers have limited density, causing significant waste of space and unable to meet the deep demands of cloud computing work. High-density blades, due to the lack of unified standards, have products from various manufacturers that lack universality and have high costs. High-density rack products use universal server motherboards, effectively reducing costs. Rack servers can achieve high-density configurations of over 80 computing routers in a standard cabinet, increasing deployment density several times compared to traditional rack servers. Rack servers also integrate management, networking, power supply, and cooling modules, providing high density to users. Therefore, high-density, standardized universal rack server products will also be an inevitable demand for cloud computing platforms.

High-Performance Computing (HPC)

In the fields of scientific computing, animation rendering, bioengineering, and CAE (Computer-Aided Engineering) at cloud computing centers, a computational model may require hundreds of millions or even trillions of floating-point calculations, which cannot be provided by a single computer due to the immense computing power required. High-performance computing has unique advantages in architecture, high-scalability interconnect networks, high-throughput communication, rapid deployment, system control at scale, and ultra-large-scale parallel computing, making it suitable for scientific computing. Therefore, efficient computing is the absolute choice for cloud computing data centers in parallel computing fields such as scientific computing and is an important technical direction for cloud computing centers in the computing field.

Massive Storage Devices

Cloud computing, in essence, needs to face numerous basic hardware and software resources; at the upper level, it needs to deal with diverse business applications and numerous end users, which will inevitably involve massive data processing that depends on a very large amount of data. Ordinary storage devices cannot meet the requirements for performance, stability, and availability in data processing. Therefore, efficient and stable massive storage devices are needed for support. Massive storage devices are the foundation of cloud computing platforms and an important direction in the trend of cloud computing technology.