The difference in Qingyun is critical to the business

We are in the midst of a long period of hype over cloud computing, which means that it is critical to have clear criteria for evaluating and evaluating different cloud options. In other words, what cloud methods should http://www.aliyun.com/zixun/aggregation/15530.html > large enterprises use to optimize their data center operations?

Typically, the cloud is conceived as a computing model that is always available, accessible, and low-cost in the air. Although this is a very ambitious cloud, companies still need to carefully select and build their cloud environment to avoid damaging their computing power, locking themselves in a single, cost-efficient environment, weakening their ability to differentiate themselves and gain competitive advantage.

The following chart provides a preliminary basis for distinguishing between different cloud computing differences. Note the location of two of the major cloud types:

-Specialized SaaS services, from server to application (even version) of the entire stack is very small change.

-the common infrastructure and Platform as a service (IaaS and PAAs), where both the server and the operating system have available storage types. To install any compatible database, middleware, or application on it

。

The private cloud is basically IaaS, the enterprise can build itself. Private cloud is the evolution of enterprise virtualization servers and storage farms to more mature instances with well-defined service configurations, product and billing capabilities, and highly automated provisioning and management capabilities.

Another technology that affects the datacenter is the engineering stack (engineered stack). This is the further evolution of computer devices over the decades-integrating standardized, well-designed, engineered components together to deliver superior performance and cost advantages.

Typically, these devices are already present in the areas of networking, security, data warehousing, and professional computing. Firewalls and other security devices are also using this approach to evolve by tightly integrating some common technology CPU, storage, os--with proprietary software, and then selling it as a packaged solution. This method of engineering stack has already entered into many fields such as data analysis, application server and middleware.

Over the next five years, manufacturers will continue to expand their IaaS and engineering stack products because they want higher margins and a more determined long-term revenue stream. Customers accept SaaS because of their ease of implementation and the potential for cost variability, and they accept the engineering stack because of its higher performance and lower cost. Whether the enterprise should choose SaaS or an engineering stack to satisfy its business example, the following considerations should be noted:

SaaS Considerations:

If it's a core business function, you have to be very cautious. Because it could be a loss of differentiation and ultimate competitiveness.

-Before signing a contract, an enterprise needs to know how to get its data back if it terminates the use of SaaS applications.

-Be sure to ensure the integrity and security of your enterprise's data in the hands of the application vendor.

Project Stack considerations:

-Be aware of where the product is in its lifecycle (old products generally do not provide lasting benefits).

-When the product is close to the end of its lifecycle, the final migration path should be prepared in advance.

For both of these products, the enterprise must avoid integrating the key business logic into the manufacturer's products. Otherwise, businesses face high migration costs once the product lifecycle is terminated, or when better alternatives emerge. There are several ways to ensure that the key features and business rules of the enterprise remain independent and modular beyond the vendor's packaged services.

Mindful of these considerations, enterprises can make successful decisions at the project level. If you are a large and medium enterprise, you should drive the internal infrastructure to the private cloud. Virtualization is just the first step. The enterprise should eliminate or minimize the custom configuration as far as possible-preferably reduce the number of its own servers below 20%. Next, invest in tools, processes, and projects that automate the configuration and management of your data center. Bigger This will also improve the quality of service.

It is important to make sure that you do not transfer too many processes to SaaS, because doing so will allow your enterprise's own utilities to become Balkan. If the stall is too large, it is expected that heavy integration costs may occur in the subsequent process, as the enterprise's functions are distributed across the data centers of many SaaS vendors. There is also the expectation of experiencing performance problems, as workloads are now running at a WAN speed rather than a LAN, and are expected to lose bargaining chips with SaaS vendors because the enterprise has lost its "inner pack" capabilities.

But over the next five years, SaaS and engineering stacks will be the cloud product mix of almost every enterprise. Powerful enterprise IT departments will use these features wisely and prudently to avoid vendor lock-in and other traps, develop their own private cloud capabilities, and retain critical intellectual property assets. With the reduction of customized configuration, we will see higher data center efficiencies. So while the outlook is certainly cloudy, the future is bright for thoughtful IT departments.

(Responsible editor: The good of the Legacy)