Use the WebSphere Creator Server V8.5 integrated Intelligent Management feature component

This article describes the related features and usage scenarios for dynamic clusters and their impact on performance.

With the growth of business scale, more and more applications are deployed in modern enterprise's IT system. Each application has different requirements for computing resources and has different amounts of traffic at different times. How to establish a flexible http://www.aliyun.com/zixun/aggregation/13748.html "> infrastructure, according to the needs of each application to dynamically allocate resources, effectively improve resource utilization, is a major challenge for enterprises."

The Intelligent Management (Intelligent Management) component is consolidated in the WebSphere creator Server V8.5, which is known as the WebSphere Virtual Enterprise in previous releases or WebSphere Extended Deployment Products. It provides a complete application infrastructure virtualization platform for middleware servers (WebSphere creator server or Oracle WebLogic server), enabling enterprises to flexibly and dynamically adapt to changes in business requirements.

Dynamic Cluster is one of the important functions. It is a kind of virtualization cluster technology, which can automatically allocate resources among different clusters according to the change of request load, thus improve the utilization of computing resources and improve the performance of application. This article will introduce a dynamic cluster and the usage scenarios for related features.

Limitations of traditional application infrastructure

In the traditional IT application infrastructure, in order to reduce single points of failure and achieve high availability of resources, organizations typically centralize a set of server nodes as clusters and deploy applications on the cluster. This cluster can be called a static cluster because its size is determined when it is created. This architecture poses the problem that it is difficult to find a reasonable way to divide the size of each cluster, using limited resources to meet the full application requirements.

Typically, each application's need for computing resources varies with time and traffic, each with its own free time. Consider the scenario shown in Figure 1: There are three clusters in the data center, and the applications deployed on each cluster vary significantly over time. Because of the limitations of the architecture, although the overall resource utilization of the data center is not saturated, a cluster is overwhelmed at every time and cannot meet the needs of all applications. The root cause of this problem is not the lack of computational resources, but the inability of different clusters to share resources to complement each other, resulting in lower overall resource utilization.

Figure 1. Traditional data Center

Modern enterprises need to build more flexible infrastructures to solve such problems. The dynamic cluster was born for this.

Using dynamic clustering to improve resource utilization and application performance

The working principle and advantage of dynamic cluster

Unlike traditional static clusters, dynamic clustering is a kind of virtualization cluster. It allows a user to make a set of server nodes into a pool of resources called node groups. Users create dynamic clusters in the resource pool and deploy applications. This approach completely separates the application from the physical infrastructure, and the application is no longer pinned to certain nodes in the cluster. Depending on the access situation of the application, the dynamic cluster will automatically start (or stop) more servers in the cluster to run the application, allowing applications to be dynamically placed in the resource pool and migrate workloads at any time to meet business needs. Through the way of sharing and allocating resources among different clusters, dynamic clustering can improve the utilization of computing resources and improve the overall performance of the application.

We rebuilt the data center in the above scenario and created three dynamic clusters, as shown in Figure 2. They automatically coordinate the allocation of resources within each node group (resource pool) based on their respective resource requirements at each time, increasing resource utilization.

Figure 2. Data center built with dynamic cluster