Grid Computing: Is It Right for Your Company?
The first step in beginning to identify whether a grid solution is applicable in your organization is to perform a detailed planning assessment. Results of the grid assessment will be used to define the most cost-effective, appropriate grid implementation plan to meet corporate goals.
The grid assessment measures the potential for practicing grid solutions within a corporation, determines whether the corporation is ready to embark on such a grid program, and defines where to focus efforts to gain the maximum benefit. The result of the grid assessment can be used as the basis for defining corporate grid goals, grid adoption strategies, the business domains where a grid might be most effective, and the basis for a grid implementation plan.
The emphasis in this article is on a business viewpointlooking at the reasons why grid solutions have evolved, and the expected business value to be gained from grid implementation.
The grid assessment is performed to help introduce a grid into a corporation. The goals of the grid assessment are as follows:
Evaluate or develop the corporation's grid strategy and the implementation of that strategy in current software projects and for various systems, groups, and end users.
Use the results of the assessment to determine the corporation's grid goals, elements of a grid program to achieve those goals, and domains in which to focus grid efforts.
Recommend actions and develop a detailed plan to implement the grid strategy.
Instituting the technology of grids across a corporation is a large, complex task, especially if the ultimate goal is to introduce grid practices at the project levelthat is, across teams, product lines, and software groups/organizations. Success requires careful planning, cooperation, and good management practices. To ensure success, a corporation needs to determine how ready, willing, and able it is to utilize a grid computing environment and what actions will prepare the corporation to accomplish its grid objectives and goals.
The grid assessment investigates both technical and management/organizational grid issues. Following are some of the important management/organizational issues to consider:
Identifying and defining basic business objects and other kinds of components
Defining guidelines and standards for core business objects (once they exist) and for creating or reengineering core business objects
Defining the organizational structure and classification scheme for the grid management structure, architecture, and utilization
The technical side includes issues such as the following:
Defining IT planning and operational support for core business objects/components
Establishing grid-training programs
Establishing the grid measurement infrastructure (grid metrics and measurements, corporate grid policy, grid service levels)
On the technical side, many issues need to be addressed in the assessment, with answers provided in areas such as these:
Availability monitoring. The grid infrastructure will contain mission-critical applications, network devices, databases, servers, and so on. The availability of network infrastructure and mission-critical applications is very important as more and more users rely on IT services for their daily operation. As the IT infrastructure becomes increasingly complex and many diverse applications are deployed, service outages may become unpredictable and undetectable due to the interdependency of systems.
Problem management. Grid problem management handles the conditions that cause users to lose full or partial functionality of any IT service. This includes problem determination, notification, resolution, problem tracking, and control.
Capacity planning/trend analysis. Capacity planning and trend analysis are important for predicting future grid performance with new architecture and services, in order to plan acquisitions to meet future demands.
Performance monitoring and response time. Performance monitoring is a process of capturing and analyzing performance data from various grid components of the IT network (including servers, applications, and network devices) to locate bottlenecks.