The best project in cloud testing

According to Xiaobian, application testing and development projects are a good place to start for those who are just getting started with cloud computing services. James Staten of Forrester Research tells you how to tell if your project fits into the cloud so it can grow faster and better.

There's really less exaggeration for cloud computing - especially when it comes to saving money for the business. Unfortunately, however, the vision of cost-saving that we have described for our advocacy seems very consistent. After countless customer surveys, the results are clearly presented to Forrester, and if you look at all of your business as a whole, the return on investment (ROI) from the cloud has not been realized because, When the advantages of cloud computing greatly reduced. Ultimately, this means testing and developing new applications in the cloud is a great way to increase business justification and to ensure that applications, Infrastructure as a service (IaaS) solutions, and IT operations protection Compatibility with the process.

IaaS (Infrastructure as a service) The rationale behind cloud computing is that those platforms provide a standardized, automated virtual environment that IT professionals can more easily use with less human intervention. Public clouds, which will also include private clouds gradually, are platforms that calculate resource consumption, which drives a different form of consumption-in which case cost savings can be achieved by minimizing resource consumption. This key concept is crucial to understanding the specific business case related to cloud computing.

For example, if you ask a test lab manager what the challenges they are facing, they may complain that developers are asking too much and "wasting" a lot of time for them to install and uninstall the test environment. But, on the other hand, for those developers who are among the most productive and innovative in a team, how can they concentrate on working in the face of so many complaints? They will tell you (and usually they will not), and they think if they do not use the resources of the labs provided by the IT operations department, they go straight to platform-as-a-service (PaaS) or IaaS (Infrastructure as a service ) Testing on the cloud (in which case they can get the resources they want in minutes, and they only pay for them when they need them) and their productivity gains ?

Of course, while some developers are migrating their work to the cloud, IT operations may worry that they are pushing the entire organization to the brink of danger. We should not stop testing and developing in the cloud. Instead, we should encourage this use, and we must give some guidance. For example, to establish a practical cloud usage strategy, or consider using a centralized resource request tool to regulate how developers use the cloud.

Here, there is a very important issue to note that not all development projects are suitable for completion in the cloud. It takes some time to define the most appropriate specifications for collaboration between the IT operations team and APM. Here are five common best practices that you can use as a starting point:

1, the deployment test can be carried out in a standard virtual environment.

IaaS (Infrastructure as a service) platform provides server virtual machines and virtual storage volumes. The PaaS (Platform as a service) cloud provides more advanced abstractions - middleware or directories for deploying applications. In most cases, no dedicated physical resources are provided.

2, the test environment to be independent.

If the development project can be tested separately, which means it does not need to integrate with the production system, then it can be tested on the cloud platform. For the cloud platform, those who do not consume the output bandwidth of the application cost is the highest. They are also less risky for the company because there is no need to open certain ports on the firewall.

3, those projects whose life cycle is less than 12 months.

Most cloud platforms are charged on an hourly basis, and are more costly than traditional hosting if used continuously for more than 12 months. So, for such a project, in the internal virtual experiment environment will be more cost-effective. This is a basic rule of thumb: if you can often make the cost of using a public cloud approach 0, then you are effectively using it.

4, those projects that will not bring new compatibility and management risks to the company.

Do not put these test items in the cloud until you think you can guarantee the compatibility of those items.

5, the use of Web services multi-virtual machine applications.

Public cloud can assign client tasks to multiple standard virtual appliances and then connect them through IP protocols and Web services. If developers want application components to interact with each other, do not expect to find a supported protocol - either a delay unacceptable, a specific network configuration, or a tightly coupled component. Most public clouds do not support multicast protocols, and clustering is often more difficult. Do not put such an application in the cloud until you can make sure that it works. However, interactions based on existing web services can proceed normally.

If accelerating time-to-market is a top priority for new applications and services, testing and development projects that meet these criteria will give developers more flexibility and business-agility To reduce the cost of using the cloud platform. It also charges in a reasonable manner because it turns long-term investments and necessary operational costs into more efficient and more flexible operating cost investments. It can also reduce the need for internal lab resources so that budgets already allocated can be used to do something more meaningful, depending on, of course, the size of the enterprise's development projects and the suitability of those projects Cloud