Those things of the private cloud

Cloud computing is a new thing in the IT field, attracting many eyeballs in IT industry, but its definition is not so clear, to say that it is not conclusive, frankly speaking, it is the definition of the following is not the same, divergent views, so contradictory. Where is the cloud? This is one of the most fundamental issues of cloud computing.

The whole concept of cloud computing began with the public cloud resources outsourced by IT business. But companies that dabble in real-world cloud computing quickly realise that most it businesses can't outsource, does that mean they don't have cloud computing at all, or are their cloud computing fundamentally private? The answer to these questions, if cloud computing is fundamentally a private cloud, how should we respond?

To some extent, most companies involved in cloud computing projects assume that private cloud is an enterprise data center architecture that replicates the data center of a public cloud vendor. When asked "What services can private cloud provide?" "When this happens, IT managers tend to answer," infrastructure is the service (IaaS). They are largely of the view that private cloud is based on virtualization technology. When asked what is the difference between a private cloud and a server-cured virtualized datacenter, most of them are unable to give the right answer.

Unfortunately, most cloud-computing providers also support this fallacy. Currently, virtually all information about building private clouds is such as improved virtualization tools and technologies. In most cases, the product strengthens the functionality of centralized resource management and focuses on a data center with virtualization technology.

Some companies have also come in contact with open source cloud computing tools such as Hadoop or eucalyptus. Hadoop created a data-model-driven cloud computing architecture, while eucalyptus almost completely rebuilt a virtual machine cloud similar to Amazon's EC2. If building a private cloud means building a cloud of agree meaning, these tools seem to provide a logical hand for this work.

Justify Private cloud computing

While these two private cloud models may be logical, companies will find themselves in the process of implementing them in conflict with their current practices and creating some economic constraints.

First, many key applications are not suitable for both modes, and often you cannot easily or effectively adapt these applications to these patterns. Because most mission-critical applications are currently running on a multitasking computer system, this is also consistent with the principle of platform-service (PaaS) service-oriented Architecture (SOA) in cloud computing. Virtualization cannot be applied to these applications, or to these applications, virtualization simply has no value, and cloud computing software tools cannot create PAAs clouds.

Second, the benefits of an enterprise implementing a specific private cloud model to replace the current IT infrastructure may not be possible at all, as it may not exist at all. Proving the significance of creating a private cloud in a datacenter is often unconvincing because of the limited overall and cost potential.

The efficiency of cloud computing data centers did not grow exponentially or even linearly with the size of the resource pools it created; instead, it grew rapidly at an early stage, followed by slow growth and finally stagnation. After examining the costs and benefits of private cloud tools, companies are finally beginning to understand that their existing data centers are at a standstill. The implementation of cloud computing will increase their operating costs by increasing the cloud management tasks based on the current load.

If the current multitasking data Center application is designed for cloud computing and conforms to the SOA PAAs Cloud model, you may already have a cloud in terms of practicality. This is what many enterprises gradually realize. The purpose of an enterprise is to run applications, not just to build expensive IT infrastructures such as data centers. If a cloud behaves as if it were a black box that hides all the details, then we can argue that most data centers are already using the cloud model that the application delivers.

The new concept of "My data center is a cloud" is useful, focusing on where private cloud technology planning should be implemented, and confirming that existing data center architectures deliver applications as efficiently as delivery services. If not, the enterprise could consider using more cloud computing tools such as Hadoop, eucalyptus, or vcloud to improve efficiency, or choose to outsource some applications to the public cloud.

When data center usage is low, implementing private cloud computing is a very necessary choice. In general, organizations will find that unordered planning for server consolidation in virtualization produces inefficient resource pools, which can be significantly improved by using private cloud tools.

If the enterprise does not have a solidified server, and some particular application is likely to be inefficient, it should be relocated to a common cloud, such as a PAAs, that uses the same middleware as the Enterprise data Center. When planning for a mixed cloud application, it is most appropriate and advantageous to treat a data center as a private cloud.

"Edit Recommendation"

Cloud computing next seven hot IT jobs next battleground: Cloud computing PK Outsourcing Cloud computing: Crisis or opportunity for data center vendors? "Responsible editor: Xin Wei TEL: (010) 68476606"