Private cloud computing platform construction: Automated management of Virtual platform

In the past few years, virtualization has been a catalyst for data center transformation by reducing capex. In the face of the resulting technical and management complexities, automated management is introduced to further optimize opex and business agility. Automation management (automated provisioning) is one of the important features of cloud computing, which defines self-service service request, quasi real-time deployment, dynamic and granular scalability.

Many organizations found that the complexity of their data centers increased rapidly as the size and type of information increased rapidly. WEB services and composite application environments also bring more complexity. The number of nodes that the IT team needs to manage is increasing, and the number of alarms, changes, and management data generated on the nodes is increasing, and the amount of work requests is increasing simultaneously, which creates a greater workload for the IT team. Use manual or ad hoc processes to keep pace with the growth of these messages.

In the cloud delivery model, cloud service providers, including but not limited to it teams, also need to deliver scalable computing capabilities based on policy and differentiate between critical business and non-critical business uptime and availability. IT management is increasingly needed to automate the consolidation of physical and virtual system management and increase the overall utilization and end-to-end resolution of IT resources. IDC believes that automated infrastructure configuration, delivery, and asset management capabilities will help cloud service providers respond more accurately and quickly to changes in applications, improve capacity planning capabilities, and effectively control costs.

Automation is a set of it strategies, not just a software or a feature. In general, automation needs to be implemented by implementing a holistic solution. For the cloud computing platform, the following are key to automating:

Monitoring

To ensure that the cloud computing platform is operating properly, the system must monitor the entire cloud computing architecture to enable system administrators to control the entire architecture in time to optimize resource performance and availability. Monitoring is usually done by monitoring software. The flexibility and high availability of cloud computing platforms can be supported by monitoring critical system resources, detecting bottlenecks and potential problems, and automating recovery in severe cases. The objects of monitoring include system hardware (server, storage, network, etc.), including software (application, database, middleware, etc.). For example, the system must monitor the real-time situation of CPU allocations, usage, and load to allow allocation to occur in a timely manner when the workload of an application changes. As mentioned earlier, cloud computing platforms usually include a variety of types, multiple architectures, and multiple brands of hardware and software that support the monitoring software of cloud computing platforms that require high compatibility and can simultaneously monitor heterogeneous hardware devices and software.

The results of the monitoring can be provided to the system administrator and can be processed automatically by the system. Monitoring software must have a complete monitoring process management function. Can complete the set performance threshold, after the valve value automatically issued alarm information, tracking the cause of the alarm, and issued the corresponding management instructions, take the corresponding operation, such as a series of work, so that the system monitoring and processing need not necessarily through the system administrator to carry out.

In addition, the monitoring function also needs to have certain forecast ability, through the historical data analysis resource consumption trend in cloud computing resource pool, forecast and avoid problem occurrence.

Dynamic application Migration

To support the flexibility and high availability of cloud computing platforms, as well as stringent service level agreements (SLAs), business applications on cloud platforms need to be able to migrate dynamically between virtual machines. For example, for resource requirements for an important application, if unexpected spikes occur, and contention for resources occurs. The system can move it to a larger server, or move other, unimportant applications to a different server and use the freed resources to meet this spike.

As application requirements change frequently, the infrastructure on which they depend must be able to adapt to new requirements in a very short time, but should have minimal impact on service levels as much as possible. This requires a simple and secure way to apply configuration changes, eliminating the need for administrators to intervene too much to reduce the cost of change management and reduce the risk associated with it.

Without a dynamic migration approach, most reallocation of resources needs to be carefully planned, performed by technicians, and lead to a certain amount of downtime, which can seriously affect service levels (SLAs).

Dynamic application migration allows for uninterrupted maintenance or change of the system. This eases the impact of the need to temporarily shut down systems or applications. Dynamic application migration operations must be performed not only when an idle (inactive) application is applied, but also when the application is providing a service (active state).

During a dynamic application migration, there is no effect on system operations or user services. For example, you can migrate a dynamic production database, which contains regular user activity, to another virtual machine without losing any data and connectivity, or having any impact on the running transaction.

Automated deployment

Automated deployment is also one of the key features underpinning the cloud computing platform. Traditional manual deployment is a time-consuming and laborious process, usually composed of several complex steps. including software installation, configuration, and the allocation of hardware resources for the software. Because the custom business application usually has the special installation and the configuration step, causes the application software deployment to become the complex process. These factors make automated deployment The key to managing these tasks on a cloud computing platform. It is only through dynamic deployment of business applications that the flexibility of cloud computing platforms can truly be realized.

The two tasks most frequently performed by cloud computing platforms are the loading and unloading of business applications. The load process installs the operating system and other software to the appropriate location on the server and configures them to make it effective. The uninstall process refers to the steps that are required to automatically recycle the server to be used as he does.

Automated deployment software requires the ability to discover and track the resources of the cloud computing platform in a timely manner, enabling accurate server settings and software deployment. Software deployment is actually a set of IT policies and processes, and automated deployment software needs to support system administrators in formulating policies, writing processes and setting configurations, and thus quickly and efficiently structuring and managing IT resources.

In addition, due to the heterogeneity of the cloud computing platform, automated deployment software should also be able to support the deployment of business applications on different hardware platforms and different operating systems.

In practical application, the function of automation management is much more than monitoring, dynamic migration and automatic deployment, and it should include automatic management of service request, resource allocation management, capacity planning management, self-service management, etc. Enterprises should build automation management platform according to their own business application.

IDC research shows that automated management can provide better and more standardized delivery services and more flexible response to changes while reducing costs for the enterprise. IDC expects more Chinese companies to virtualize and automate infrastructure operations and resource allocations-creating a private cloud computing environment. At the same time, with the increasingly wide application of smart assets in various industries and processes, organizations will have the opportunity to extend the experience gained in IT asset optimization and automation to the process of enterprise asset optimization.

(Responsible editor: The good of the Legacy)