|
Service execution is the process for delivering operational services to the IT infrastructure and the enterprise. In effect, it is telling applications and systems what workloads to run, when to run them and where to run them throughout the organization in a consistent, predictable fashion. While the service execution process is straightforward, it is exceptionally difficult to manage. The right management tools should be able to enable business policy–based responses to planned and unplanned changes, in order to optimize workload velocity and IT resource utilization.
Traditional management tools such as job schedulers, load balancers and cluster managers can automate some of these activities, but they leave IT organizations without the ability to automate, dynamically manage and optimize service execution activities end to end, across the enterprise.
These challenges can be effectively addressed by dynamic, end-to-end workload automation solutions that provide a virtual point of control to implement standardized service execution processes and support a service oriented architecture (SOA). Solutions should be designed to respond quickly to service demands and changes in the IT environment, balance computing costs and service levels, and improve utilization of IT capacity.
Review current challenges in managing service execution
As shown in the figure on the next page, service execution is the process for delivering operational services. The overall process represents the “central nervous system” of the enterprise. In effect, service execution tells applications and systems what workloads to run, when to run them and where to run them throughout the organization.
While the service execution process is straightforward, it is exceptionally difficult to manage. Management tools must be able to:
- Support business processes and policies.
- Enable planned changes to business processes and the IT infrastructure.
- Adapt to unplanned incidences in the IT infrastructure.
- Help maximize workload velocity.
- Optimize the utilization of IT resources.
- Adhere to stringent service level agreements (SLAs).
- Ensure adherence to compliance and governance requirements.
Additional difficulties arise from trying to manage complex, heterogeneous applications and systems as well as processes across organizational silos. Finally, IT organizations — already limited in resources — must somehow deal with a growing number of mixed, interdependent and often unpredictable workloads that need to be scheduled in real-time, near-real-time and batch modes.
Historically, enterprise-wide service execution processes have not been well defined. Furthermore, traditional management tools such as job schedulers, load balancers and cluster managers have enabled IT organizations to only manage workloads for individual IT resources, or a cluster of resources. These traditional management tools ultimately leave IT organizations without the ability to automate, dynamically manage and optimize service execution activities end to end, across the enterprise. This paradigm has led to the use of multiple workload management environments within the enterprise. As a result, IT organizations find it challenging to:
- Provide a virtual point of control to implement standardized service execution processes.
- Implement standardized, scalable workload management activities or processes.
- React quickly to changes in the IT environment.
- Properly balance computing costs and service levels.
- Optimize workload plans and choreography to increase workload velocity and improve utilization of IT capacity.
- Provide adequate reports on overall workload performance across applications, systems and organizational groups.
Recognize the consequences of inadequate workload management
The lack of efficient management for service execution can result in significant problems at both the technology and business levels.
Consider this example: a Web services invocation triggers a stream of dependent, real-time and batch workloads. These workloads span enterprise resource planning (ERP) applications and databases running in open systems, as well as messaging middleware running in mainframe systems. With individual scheduling environments for each application, middleware and system, it is difficult to understand and optimize workload processing performance across the entire service execution process. It is also difficult to monitor workloads by exception and to proactively address and understand the impact of workload errors from an end-to-end standpoint.
These technical problems can lead to a slowdown or degradation of business processes — and that can result in customer frustration, customer churn, complaints from business partners, damage to the brand and company image, lost sales and other consequences that impact the bottom line and reduce the overall competitiveness of the business. These problems can also lead to inefficient use of existing IT resources, creating a strain on some resources and the underutilization of others.
|
