Breaking cloud barriers: methods and misconceptions during cloud deployment

According to IDC data, the Global 2011 Cloud services Market size of about 28 billion U.S. dollars, is expected by 2015, the global cloud services market size will reach 73 billion U.S. dollars, an annual growth of 27%, is the traditional it growth of about 5 times times. And in China, in addition to 5 cloud computing pilot cities, 13 provinces or cities issued a cloud computing strategy, according to IDC estimates, if all implemented, the input of optical server will reach 10 million, the total investment amount will reach 270 billion yuan.

Six misconceptions in User cloud deployment

Lured by a vision of cloud computing, more and more companies are brewing to move to the cloud. What is the enterprise's understanding of the cloud? In the process of cloud migration, what are the main problems that enterprises are facing at present? How will the cloud develop? In response to these problems, IDC in the global scope, for more than 150 enterprise CIO conducted a survey, the results show:

Responding quickly to new business requirements is the main reason why companies consider deploying cloud infrastructure, and second, to make more efficient use of IT resources and to standardize it processes. In addition, costs are another driving factor in the enterprise's consideration of deploying a cloud-computing infrastructure, with about 40% of CIOs thinking that reducing IT spending is their intention to move into the cloud.

IDC also conducted a survey of the final whereabouts of the enterprise IT budget, which showed that the IT budget for the next two years was more invested in hosting private cloud and local private cloud, rather than cloud computing services. Because most of the companies currently considering cloud deployments are in large and medium enterprises, they are more likely to build their own infrastructure, such as building their own data centers, or building private cloud architectures within managed data centers without using Third-party cloud computing services. The concrete results are shown in the following illustration:

As shown above, in the next two years, IDC expects the traditional it model to drop dramatically, now that the traditional it accounts for 49%, but after two years, it is expected to fall to 37%, the local private cloud construction will be favored by enterprises, the proportion of its IT budget will be significantly increased, is expected to be from the current 13% Rise to the next 19%; the managed private cloud will also have a significant increase, from the current 11% percentage to 15% two years later, and public cloud computing services in two years.

Although many companies are attracted by cloud computing, they want to be able to achieve cloud computing ahead of time. But how to plan and implement concretely, there is no clear road map at present. How do you move from a traditional infrastructure to a new cloud computing infrastructure? IDC's survey of more than 150 corporate CIOs found that the majority of CIOs (more than 60%) still had bigger problems with how to implement them, and nearly 40% of respondents said they were not very clear, even when the vast majority of respondents chose the least "management services". As shown in the following illustration:

The IDC survey found that the problem was not only this, but also in the deployment implementation process that the user might involve, from assessment to strategy consulting, to providing a roadmap of cloud computing, to implementation, from virtualization to automation, to integration, including the management of changes after the implementation of the cloud computing infrastructure, Throughout the implementation of cloud computing, there are a series of problems and misunderstandings. By summing up, IDC believes that there are the following six main misunderstandings:

Re-implementation, light planning and consulting.

According to IDC, assistant vice president of China, 武连峰, in the 150-plus CIO of Chinese enterprises visited by IDC, they did very little in their early assessment, planning and consulting work when implementing cloud computing. The implementation process inevitably becomes "stones", encountering more difficult problems.

Heavy hardware, light software.

When most CIOs talk about cloud computing, the first impression is to buy hardware such as servers, storage, and networking. Cloud computing usually means consolidation, virtualization, and automated deployment of IT resources, most of which need to be implemented by software and middleware.

Focus on server virtualization only.

It is true that the server virtualization technology has brought many benefits to the users, it also affected the cloud computing, many people believe that the basis for cloud computing must be virtualized. But virtualization here is not just about server virtualization, but about the full range of virtualization, including servers, storage, and networking.

Heavy concept, light nature.

Cloud computing first originated in the United States, the core is to reduce costs. But now many companies are spending a lot of money on the budget, but they don't reap the benefits of satisfying their expectations. Many CIOs regret that cloud computing does not cost less, and instead spends more money than it does with the original idea of cloud computing.

Heavy Google, Amazon model.

The rise of cloud computing is not a "relationship" with Google or Amazon, and Amazon's "AWS" model has actually become the de facto standard of IaaS. Many companies have talked about cloud computing's first thoughts, such as Google, Amazon's internet model, ignoring the needs of the enterprise itself and other cloud architecture patterns.

Re-innovation, light transplant.

Moving to the cloud does not mean applying innovation, Mr. 武连峰, and it is more and more important to migrate traditional business applications to the cloud in the process of migration. But in the research process of IDC, many enterprise CIOs think that the new cloud model needs to apply innovation, but ignore the migration of traditional application.

How to migrate to the cloud smoothly

These issues mentioned above, are the IDC company launched in May this year on cloud computing infrastructure Construction Research project, concluded that most of the work of this research has been completed, and based on the results of the above data and conclusions. To address these issues, IDC will combine the global cloud computing experts to carry out framework planning, and in the domestic more than 150 CIO and IT managers to communicate, to sum up our country in the implementation of cloud computing infrastructure Roadmap, the report is expected to be officially published in June.

Although the report has not yet been officially published, but 周震刚, a senior research manager at IDC's China Enterprise-level systems and software Research department, revealed some of the details that the company's traditional IT systems are in the process of transforming into cloud computing, and that its importance will go through four steps, including consolidation, virtualization, automated deployment, and open APIs, As shown in the following illustration:

Integration is to integrate the existing IT resources in order to achieve the goal of simplifying. Virtualization is the centralization and allocation of IT resources through the adoption of virtualization technologies. Whether it's consolidation, virtualization, or automation and open APIs, this is a more mainstream cloud migration step. But IDC believes, in this, still need to pay attention to several misunderstandings.

"Cloud computing is not necessarily a distributed settlement"

According to Mr. 周震刚, in the user practice process, there are usually users that cloud computing must be distributed computing, and that cloud computing is a distributed computing, through the Distributed Application task to split into several different applications to parallel processing. In fact, it is a very one-sided view that cloud computing is actually a mobile model that provides computing resources and software as a service to end users, rather than a computing model.

Generally speaking, cloud computing is a business model, not just a technology. Distributed computing solves the problem of reliability through distributed computing, but we can see clearly that distributed architecture is reliable through distributed processing at software level. To measure the overall cost, the overall cost of many vendors running public cloud computing services is still high, though many open source software is used to reduce costs, but software developers account for more than 90% of the entire company, a shift in costs rather than elimination.

"Cloud computing is not equal to X86 architecture"

Many people think that the deployment of cloud computing can only deploy the X86 architecture, in fact, this is not true. The user's different business and different types of hardware requirements are actually different, so in the current situation to deploy cloud computing, heterogeneous platform is still there. To break the traditional framework to implement a new X86 framework, 周震刚 that it is too radical. And warns that this approach has a big impact on the traditional infrastructure.

IDC believes that the future of cloud computing is a very large pool of resources to achieve the core applications, things processing, databases and so on the integration of RISC architecture equipment, but also to achieve file printing, as well as the more low-end application of more open architecture, but also in external applications and collaborative applications more appropriate X86 architecture.

"Virtualization is not cloud computing"

Some users believe that the deployment of virtualization has completed the cloud computing architecture. In fact, deploying virtualization is just the second step towards the cloud, only virtualization without an automated management platform, still requires the input of managers, cloud computing does not save the cost of people, instead of a large number of people involved, not care about the automated management of virtualization, so that the cost of cloud computing has increased.

And virtualization is an enterprise-level, full virtualization, not just servers, but also storage, network devices, and so on, all of which need to be fully virtualized, integrating all the cloud infrastructure resources to achieve virtualization. In addition, virtualization is not step, to consider the enterprise-class application extensibility, availability, security and other aspects of the main points.

According to the research data provided by IDC, 60% of all users surveyed have deployed virtualization technology, but only 30% are actually automating management. The overall virtualization of the entire architecture, and ultimately the ability to adapt the system, the users of the service in the entire access process accounted for only the proportion of percent, pitifully small. So IDC suggests that enterprise-class virtualization is the cornerstone of building a cloud computing architecture, without virtualization it is difficult to establish an infrastructure for dynamic allocation of resources, and how to manage virtualized environments and create virtualized environments; Furthermore, technical content such as dynamic management, lifecycle management virtualization management needs to be considered; Realize the virtual management of the Self Service, adaptive, this is the final result.

(Responsible editor: Lu Guang)