Building cloud computing is not a goal cloud application is the core

What are some of the pitfalls in building cloud computing? is the Google model of cloud computing right for all companies? What kind of virtualization meets the needs of cloud computing, and in the following, you may find the answer.

The experience of the cloud

Now many companies mention the successful model of cloud computing will be associated with Google, and then immediately think that the distribution is cloud computing, but Google and Amazon's business applications and technical strength and their own enterprises are different, once deployed distributed cloud platform, resulting in two island-style architecture, resulting in the waste of resources, Despite the rapid deployment of cloud computing, it did not achieve the final results. The construction of cloud computing should be tailored to the different nature of the enterprise and the business application, and should be promoted to a strategic level.

When you bring up cloud computing, you will talk about virtualization, so why is virtualization so important? Virtualization can improve the efficiency of equipment, improve the flexibility of equipment, and obtain more resource utilization. While most users are already aware of the importance of virtualization, it is important to note that the cloud computing platform requires enterprise-class virtualization, and enterprise-class virtualization is the foundation of cloud computing.

Regardless of x86 architecture or RISC architecture, no architecture in the world is necessarily the best, as long as the business architecture for the enterprise should be the best, and any architecture should be combined with the business. Different workloads in an enterprise are placed on different architectural platforms depending on their nature, running on x86 architectures for general business applications, and key businesses deployed on UNIX platforms. Finally, even though the cloud computing architecture is not necessarily unified, it is to achieve unified management. In this context, IBM Starter Kit for Cloud (SKC) can transition your virtualized environment from "Cloud ready" to a real "cloud" environment, SKC a cross-platform, fast cloud service management software that provides seamless support not only for the power platform, It also provides support for x86 and even z that users already have. SKC provides integrated support for x86 KVM with power, unified deployment engine, unified management interface.

A third platform based on cloud computing has arrived, according to a report by IDC, the China Cloud Infrastructure building Guide. What is the third platform? The first platform is to host as the core, the second platform to PC as the core, LAN, the Internet as the core, the third platform to mobile devices and applications as the center, is actually a human center. Cloud services, large data, Business Analytics, mobile broadband socialization, these four aspects as the core of the third platform. In this case, the biggest foundation is the cloud, the cloud at the bottom, the higher the social mobility, the greater the amount of data produced, the greater the need for IT infrastructure. So cloud computing is an important way to realize the wisdom industry, and cloud infrastructure is an important foundation.

For the industry enterprises, the Enterprise service is a big trend, especially the service trend of Chinese manufacturing enterprises, many manufacturing enterprises have been to the service transformation, but is no longer our impression of the traditional manufacturing enterprises only production products, but to the user or indirectly for the provision of services. Therefore, the adoption of cloud computing can enhance the core competitiveness of enterprises. Enterprise business Change and transformation, need IT system to do a good support, these support must have the background cloud computing, the need for data integration.

There are six misunderstandings in building cloud computing in enterprises

武连峰, assistant vice president of the IDC (China) industry Research and Consulting Services department, spoke about six myths about how companies are building cloud computing at IBM Power Cloud's new generation of enterprise Infrastructure Cloud Media conference:

Myth One: Emphasis on concept, light nature: Users tend to focus on how much hardware and software products or services to buy, ignoring the nature of cloud computing is to improve server resource utilization, cost savings and how to provide instant service.

Mistake two: Re-implementation, light planning:

Myth Three: x86 architecture is cloud computing:

Misunderstanding four: Heavy hardware, lighter than pieces: Many companies mention cloud computing, buy a server immediately, buy storage, if the cloud does not have a good infrastructure software, no good system management automation software, cloud computing hardware can not play its role.

Myth Five: Heavy virtual, light management: Most enterprises are only to achieve local virtualization. The nature of cloud computing should be a full range of virtualization, which requires good management to automate it and improve the overall it architecture.

Myth Six: Re-innovation, lightweight porting: Most users deploy cloud computing to deploy new applications, ignoring the porting of existing applications.

Based on the above six misunderstandings, IDC recommends that the implementation of cloud computing infrastructure should go through the following three periods:

First period: One is planning, the other is preparing

Priority should be given to two points in planning, one is to clarify the purpose of cloud computing for enterprises, and the other is to ensure the participation of top management. From the perspective of the entire infrastructure of cloud computing, promoting business is more important than cost savings. Cloud computing is not only the change of it, but also the change of the whole architecture of it, which is closely related to the change of the business, so it must have the participation of the highest leadership and will be more handy in the process of implementation.

In the preparation phase to identify what the enterprise IT applications, the enterprise's own business analysis, the different business will adopt a different system architecture implementation. For the core applications of the enterprise, high delivery costs, the infrastructure for an enterprise-class private cloud; some data analysis, including collaboration, is better suited to a converged cloud infrastructure, and applications related to consumer automation are ideal for open cloud computing architectures, like e-mail or Web sites, which are more suitable for public cloud outsourcing.

Second period: implementation and deepening

Virtualization is the beginning of cloud computing, but it is important to establish an enterprise-class virtualized resource pool. Create a pool of resources through virtualization, where security is the most important technical feature for users, while flexible trunk and scalability, with hardware performance behind security.

Full range of virtualization, including servers, storage, networks. These three virtualization, one is the virtualization of computing power, the second is the virtualization of storage capabilities, the third is the virtualization of transmission capabilities, so that the server's computing resources, storage resources, network bandwidth resources will be fully utilized, there will not be a bottleneck. But the reality is that users buy a lot of servers, but if you do cloud computing, you need to buy some virtualization software to support the entire cloud computing platform infrastructure, and it turns out that buying virtualization software is even more expensive than buying a single server. Finally, we should use a stand-alone system or even the cluster of hardware to build up, in the operation of the use of the process of many problems. One is from the performance and resource coordination did not put these servers do a good resource utilization. Second, from the perspective of automation, did not do very good automation, including the process of resource allocation, many need IT staff participation, greatly reduced efficiency.

After virtualization is completed, it is important to automate the management and resource automation, and finally realize the overall resource scheduling, including real-time monitoring of resource pools, including lifecycle management, image management, rapid deployment, dynamic migration, etc. In this respect, the difference between domestic and foreign users is very large, domestic users in this aspect of investment than in foreign countries to be stingy, in fact, if not do a good job of management automation and resource automation, do a good job of resource pool matching scheduling, in fact, the entire cloud computing infrastructure equipment did not play well, the overall it efficiency did not improve.

From virtualization to automating the cloud-deepening process, is divided into 4 stages, the first phase is virtualization, according to IDC Global research results show that the server virtualization rate has reached 64%, the second phase is the integration of virtualization and automation processes, the number of users have 32%, the third phase is the virtualization, Automated processes and converged infrastructure, including networking, storage, and less energy from green and overall efficiency, are still low, only 3%. Phase IV is an ideal situation, through virtualization, automated processes to the fusion architecture, and finally self service, adaptive, which greatly reduces the pressure on IT staff, while enabling it to better integrate with the business. This part of the user is only 2%.

Third period: application and Management

Now the company to build cloud computing seems more blind, has forgotten the original purpose, but only for cloud and cloud, so that the cloud is not the goal, in the cloud running what application is the key.

Workloads are the core of cloud computing or data centers, and the application of a public cloud or building on a private cloud based on a private cloud infrastructure is primarily a legacy application. In addition, enterprises are more inclined to deploy the core business in the private cloud, which is consistent with the overall security of the enterprise or the current application of the enterprise and the future development of a fusion point of view.

Cloud computing is by no means only supporting new applications, and porting the original application is also critical. Similarly, for domestic users, it should also provide two development of APIs to do cloud computing infrastructure, mainly because now most domestic users also need Third-party developers to customize the development of applications based on enterprise needs.

(Responsible editor: The good of the Legacy)