Four key steps to successfully implement cloud computing

For the enterprise, to achieve the successful implementation of cloud computing, to achieve the expected business goals, take a pragmatic step is the key. Over the past few years, platform has explored with many of the world's successful cloud computing clients and partners, summarizing the four key steps to successfully implement cloud computing: survey Learning, using scenarios, validating experiments, and putting into production.

Cloud computing is a new thing and there are not many successful cases, so companies need to investigate and understand the technologies, products, benefits, and risks involved when deciding to implement cloud computing. After investigating the study, the enterprise should make clear how to build the cloud, what is the goal to achieve, and what is the plan to implement, the next thing to do is to actually build the cloud environment, but not at this time to invest in large-scale production applications, but from the selected infrastructure and applications began to carry out experimental projects, Verify the applicability of the cloud, the end is to put into actual production use.

Step One: Survey learning

To migrate from a chimney-like server-centric datacenter to a cloud data center, the first thing to do is to investigate and learn from your business, to learn about the cloud, to identify the benefits of the cloud, and to turn to the cloud. In this connection, there are three areas that require special attention.

One is to define a concept: cloud computing is not just virtualization.

When companies consider cloud computing, the most common confusion is the relationship between "virtualization" and "cloud computing". Although many cloud environments are built using virtualization technology, virtualization is not synonymous with the cloud. Cloud computing needs to integrate and deploy large-scale resources, it often needs virtualization technology, but also beyond virtualization technology, in the cloud computing environment, both virtual machines, as well as physical machines, all resources need to be centrally managed and scheduled. As the Forrester senior analyst James Staten said, "Virtualization is a step towards cloud computing, but achieving virtualization does not mean cloud computing."

Second, understand the core technologies and products needed to implement cloud computing.

Cloud computing is a new technology, but it is not all the aspects that are involved. For example, cloud computing resources and applications are not new: hardware resources or X86 servers, small machines, mainframe, storage equipment, and applications are the original business applications, but is to deploy in the cloud inside the run.

The new cloud is reflected in the management of the cloud. We can use simple supply and demand to simulate cloud computing: resources are "supply" and applications are "begging". The biggest difference in cloud computing is the emergence of a new "stack" between supply and demand, compared with traditional computing patterns, this is the cloud management software, it should assume the entire cloud environment management responsibility, on the one hand, to resource-oriented, the resources for the management and deployment of automation, on the other hand, to face the application of resources to meet the needs.

The core of cloud management platform is automation, can intelligently perceive application, resources and dynamic, flexible deployment. For cloud management software, another issue that must be considered is "open", because most cloud environments are heterogeneous, this heterogeneous is not only hardware resources, but also include software resources such as different operating systems, different virtual machines, etc., which requires that the cloud management platform is an open architecture, Able to tolerate and integrate a variety of heterogeneous hardware and software resources to support different applications, only in this way, customers will not be a certain manufacturer locked into the new "cloud chimney."

Third, understand the changes that cloud computing brings to people and organizational architecture

This includes three aspects: first, leadership and end user support. The implementation of cloud computing cannot depend solely on the efforts of the IT department, but also on the strong support of the enterprise leadership and the involvement and cooperation of the business unit, followed by the change of IT management style and organizational structure. In the traditional it architecture, the IT department's management model is also chimney-style: Server administrators do not go to the storage, the corresponding, storage administrators are not concerned about the state of the server, and the implementation of cloud computing means that the IT department from the cost center into a service center to provide comprehensive support for business applications. In this case, the change in the organization structure and management process of the IT department is unavoidable and the knowledge skills of the IT staff need to be improved. Finally, you need to understand the changes in cloud usage and operating patterns, including security, service levels, and billing.

Step two: Use the scene

After a deep understanding of cloud computing technologies and products, companies need to start from their own business and determine what applications to start with: Starting with java/web applications or starting with Business analytics applications, updating the entire infrastructure, or starting with a development and testing environment Whether to migrate the original High-performance computing environment to the cloud or to rebuild the new cloud environment. These business scenarios can be used as a starting point for the cloud, and the key is to make decisions based on the actual needs of the enterprise.

Next, you need to assess the benefits of cloud computing: saving money or providing a quick response to your business? The benefits of cloud computing vary from person to person, but there are some basic criteria to measure the return on cloud computing. The first is to strategically measure the impact of the cloud on it and the benefits to the enterprise; The second is the traditional ROI, such as how many servers are reduced, how many managers, and so on, and the risk assessment is very important.

Step Three: Experimental verification (POC)

After completing the first and second steps, the preparation for cloud computing is almost complete, followed by implementation. In this process, the following three issues require special attention:

"Focus on customer needs, it should interact with business units to understand the needs of the business;"

"Do not good wedding workers, the experimental project is not large, but to practical use;"

"The experiment of cloud computing as a project, need to have special and specialized resources, do not temporarily redeploy personnel and resources."

Generally speaking, the experiment verification of cloud computing need 4-8 weeks, if too long, the allocation of resources is insufficient, the condition does not have, need to reconsider. After the experiment is completed, the final assessment is important, including whether the implementation of cloud computing meets established objectives.

Step Four: Production use

After the first three steps, cloud computing can be put into production. At this stage, we need to take the "small step" of the way, step-by-step, do not "Great leap forward", do not tumble. For example, starting from the provision of self-service services, and then the application of expansion, flexible calculation, and finally to the billing stage.

In a cloud environment, it serves as a service, so it needs to have a comprehensive cloud service catalog that lists the various services, service levels, and billing standards that the cloud can offer, so that users can choose the services they need. The last thing to do is to gradually expand the scope of cloud applications, the gradual migration of the old system to the cloud environment, "This goes up", and finally realize cloud computing.