What is workload management?

System management typically focuses on monitoring the availability of systems. While system availability is certainly important, it often neglects the complexity of modern systems and computing environments, such as partitioning capabilities, utility pricing resources, and clustering. System availability also neglects the importance of the applications themselves and of the dynamic between (1) application performance as seen by users and (2) resource management as seen by system administrators and investors in computer resources.

To best serve financial investments and administration costs, applications and workloads must be consolidated on fewer servers and resources must be dynamically allocated as needed to provide expected performance to high-priority applications. In addition, reserve capacity should be deployable automatically to enable customers to pay for what they need only when they need it and to make the resources available to workloads that have the greatest demands at that time. In clusters, higher utilization of resources can be maintained by managing failover of workloads among servers and partitions.

To best serve end users, an application not only must be available but must also provide its services in a timely manner with expected and consistent levels of performance. For example, if a task is generally expected to complete in less than 2 seconds, but takes 15, the service to the end user is not adequate. Thus, system and application availability do not guarantee the timeliness of a service.

Workload management ensures a service’s availability through service-level management. This type of management is based on the following components:

Now to explore these components in more detail:

IT service management
A strategy driven by business requirements to provide, and in many cases guarantee, levels of service from IT (Information Technology) to end users. The needs of the end users are the primary driver for the development of the IT infrastructure. The benefits of defining this strategy are higher return on investment in IT expenditures and also the proper setting of expectations for all organizations involved.

Service-level agreements
Documents defining various service levels IT is expected to deliver to the end user. These documents focus on individual applications and the service levels required by the application users. SLAs can also be used for budgeting, planning, and control purposes.
The following steps outline how to define a service-level agreement:
1. Establish an inventory of system resources (people, CPU resources, disk space, and so forth) that serve the end users.
2. Determine how much of the inventory of system resources is currently being consumed to support the present set of applications.
3. Determine what the end users require to maintain the status quo if they are already receiving acceptable service.
4. Determine what the end users require to improve their current service if they are not receiving acceptable service.

Service-level objectives
Objectives, derived from SLAs, explicitly describe the expected utilization, availability, performance, security, accuracy, or recovery. They can also describe expected response time, job duration, or throughput.
SLOs lay the foundation for developing the actual utilization and performance metrics that IT management must collect, monitor, and report to determine whether the agreed-upon service levels are being met.

Goals
This data is used to ensure, and determine whether, an SLO is being met.

What is workload management?

Technical documentation

» Table of Contents

» Glossary

» Index