How to build a cloud factory? The cloud transition problem to develop a suitable framework

Source: Internet
Author: User
Keywords Work these servers offers is

The continuous efforts of mankind seem to be lagging behind the change in IT industry. While industry standards for it are essential to drive change in the IT industry, the power of cloud computing to drive change in it is surprising. Recently some companies are deploying "legacy systems" for the introduction of virtualization, and it is worth emphasizing that change is the only constant for data centers. Careful planning is essential to deal with this major change involving the entire enterprise workload-hosting model.

To meet this challenge, it is important to develop an appropriate framework for the issue of cloud transformation. This involves many aspects, including the agility of the cloud, flexibility, transparency, and the interests of end users. Many of these are in the final analysis related to the new specific infrastructure requirements; http://www.aliyun.com/zixun/aggregation/18871.html "> Self-Service Portal's use hosts new applications or host transient processing requirements. Agility is often a beneficial experience when hundreds of thousands of existing workloads move clouds infrastructure. In fact, the opposite is often the case, because the cloud presents a higher standard of demand (for example, qualifying the scale of the catalog and the choice of software), and actually migrating existing physical and virtual servers to the cloud model is quite difficult. In other words, in terms of the expansion of the new workload, it looks less agile in another way.

This is the source of the concept of cloud factory. In the process of industrial production, the factory is the epitome of measuring productivity. In spite of some efforts to prepare for processing machinery, it is essential to provide effective inputs and outputs once the company is to be formed. Migrate to the cloud infrastructure by adopting a common approach, the appropriate design to provide repeatable results, and the Organization's ability to devote the least amount of time and effort.

Under the guidance of this concept, it is important to further elaborate the appropriate design. Many companies can solve such problems from the grassroots, using spreadsheets and the ingenuity of smart people to determine action. The problem with this approach is that it rarely touches on the main point-the answer is really accurate, mainly because the problem is too complex. Migrating workloads to the cloud is related to many aspects, such as handling large amounts of historical data, analyzing configuration information on servers, migrating applications, scaling and software heap layers of modeling targets, enforcing enterprise and regulatory requirements, complying with service level agreements, and data protection rules. The use of spreadsheets is not very good, roughly the same way, the enterprise accounting platform is not suitable for the use of spreadsheets. Even if they are tricked into giving decent answers to this simple environment, they may not generate reports to satisfy shareholders, managers, engineers, operators, etc. All of these personnel need important and detailed decisions to ensure achievement and risk minimization.

The analysis that focuses on the cloud migration list concludes that a key concept connects all of these. This concept is a strategy that represents the basic rules including how to manage the workload, where they should go, how much resources should be apportioned, etc. There is no appropriate strategy; The hosting decision leaves the practitioner to perform the migration; There is no plan whether they did the right thing. Cloud infrastructure planning and management do not have the right strategy to try to fill out a tax application without instructions-too many variables are getting the right strategy.

After understanding all these concepts, cloud factories became clear. It divides the problem into a series of logical steps, all of which are related to data. For example, target models, planning, and management policies. In order to fully determine how the process of automation will evolve and how large. Here are the steps to build a factory:

1. Candidate Qualification: This process determines whether a particular workload is appropriate for a particular cloud environment. After qualitative and quantitative analysis, the actual candidate qualification is selected from the workload to better implement the following steps. Examples of quantified standards include: maximum I/O rate, context switch limit, maximum CPU, memory size, and so on. Qualitative criteria include: data sensitivity, service level agreement needs, backup strategy, and other factors to consider. It takes all these factors to develop a strategy to quickly make an accurate assessment.

2, Factory Size: The next step in determining the eligibility criteria is to host the cloud to the most appropriate level and utilization patterns. It also depends on policy, including how much remains to be considered and the level of target utilization. Specific details of the actual required instance size and the level of engineering utilization in the cloud environment. Explicit usage criteria are a critical step, and the conversion of CPU utilization from the current environment to the cloud environment depends on the relative speed of each CPU.

3. Balance load: This step focuses on migrating load balancers and clusters. Because the cloud environment offers different sizes to choose from, it can also provide more advanced "resilient" features. These one-to-one server conversions are not always satisfying. To be successful at the lowest cost. This result should be combined with the general scale to provide a complete plan.

4. Draw the Software heap layer: This step takes into account the operating system and software configuration of the source server and configures them in the "closest" cloud. This is an effective standardized analysis because only limited software option settings are available in the cloud directory. As an IaaS, this step is typically limited to os--level configurations, matching existing server operating system properties, operating system virtual machines, and other services on the cloud. (usually few in the list) as a service platform, this step also includes consideration of the actual software inventory and installation of the application. The result may be that the X server looks like an IIS V6 server, but it can be seen differently from the standard image by providing not only the best heap layer deployment, but also a fix list to reduce the risk in the implementation process. This is the key.

5. Reasonable arrangement: After all this is finalized, the next step is the internal cloud environment problem. The workload should be placed in the infrastructure, in effect controlling the cloud environment. Because most of the cloud is based on the virtual cloud environment, the key to adapting to the new virtual machines is to optimize the utilization of server resources. This step seems to be somewhat like placing the workload in a virtual environment (which is often similar to the size of the available server for Tetris), but the excessive use of policy has a significant impact on the configuration results. If the policy is to strictly reserve capacity for each cloud instance, the cloud environment will be very safe but relatively inefficient, with a fairly low workload (think about playing Tetris). If the policy is excessive use of resources, high-end customers may dispute higher risks. If they ask for unexpected requirements outside the cloud environment, the result of higher densities may significantly reduce costs (think of the small volume required to put Tetris tightly together).

6. Exception handling: Looking back at the first step, the usual application components or business services in the cloud hosting may not be appropriate. In these cases, it is necessary to evaluate other hosting to determine the processing method. Because there is usually a prioritization in terms of managed options, this step involves whether the rejected workload system eligibility violates the established ordered hosting strategy. These strategies include using custom cloud allocation instances, using dedicated cloud servers, hosting in virtual environments, using dedicated blade servers, using dedicated rack servers, or leaving work alone (when necessary).

Only methodically implementing these steps can quickly and exhaustively plan the cloud migration process. Through data-centric and policy-driven approaches, there are fewer mistakes and fewer rework. The owner of the application and other shareholders confidently achieve the goal. This transparency, combined with detailed specifications and implementation details, can quickly accelerate cloud initiatives. This not only reduces time but also enables the IT industry to keep up with the pace of technological innovation.

(Responsible editor: admin)

Related Article

Contact Us

The content source of this page is from Internet, which doesn't represent Alibaba Cloud's opinion; products and services mentioned on that page don't have any relationship with Alibaba Cloud. If the content of the page makes you feel confusing, please write us an email, we will handle the problem within 5 days after receiving your email.

If you find any instances of plagiarism from the community, please send an email to: info-contact@alibabacloud.com and provide relevant evidence. A staff member will contact you within 5 working days.

A Free Trial That Lets You Build Big!

Start building with 50+ products and up to 12 months usage for Elastic Compute Service

  • Sales Support

    1 on 1 presale consultation

  • After-Sales Support

    24/7 Technical Support 6 Free Tickets per Quarter Faster Response

  • Alibaba Cloud offers highly flexible support services tailored to meet your exact needs.