In-Depth
Accelerating Agility through Workload Automation
By using event-driven rules, workload automation can trigger work based on nearly any event, helping IT optimize finite resources so critical workloads run reliably.
by Denise Kalm and Keith Woodward
Today’s IT organizations face multiple challenges when managing complex workloads with finite processing resources. Organizations must optimize these resources to ensure that critical batch jobs run reliably and that multi-platform, multi-system resources are coordinated so business transactions and services work seamlessly. To fulfill these requirements, IT managers must take a more sophisticated approach to workload automation and invest in innovative software technology.
In the early days of mainframe computing, single-processor systems easily completed their allotted work during overnight batch windows. As multiprocessor systems, parallel Sysplex emerged and the mainframe could run a greater number of increasingly complicated jobs, completing those jobs on schedule became more difficult. At the same time, the batch window shrank to almost nothing as organizations faced increasing demands to process transactions around the clock. As a result, the first job scheduling tools were introduced to ensure that batch work was completed on schedule by grouping similar workloads and running them at predetermined times.
With the proliferation of newer computing architectures, job scheduling tools have evolved from simple time-stamped scheduling utilities to dynamic, event-based automation managers. For example, workload automation tools connect workloads with the business services they support, enabling IT to manage workloads as integrated business services rather than disparate IT components.
Real-World Benefits
To understand the practical benefits of this more sophisticated approach to workload automation, let’s consider a company that fully intends to standardize on a common set of platforms and applications but still develops a heterogeneous enterprise computing environment over time.
Let’s say the company decides to standardize manufacturing operations on SAP running on z/OS. While the company is likely to convert most processes to the new system, it may also choose to maintain a familiar, homegrown Solaris-based tracking system.
Suppose this company later merges with another company that uses Oracle on HP-UX. The finance department finds no reason to convert this Oracle system and it is introduced wholesale into the environment. Later, the customer service department decides to invest in a new CRM solution such as PeopleSoft on Windows. Meanwhile, the company standardizes on Lotus Notes for messaging, even though one subsidiary is already using Microsoft Exchange Server. As customers interact with the company electronically, all of these systems must work together to receive and process orders, determine pricing, track inventory availability, set shipping dates, answer customer complaints, and alert key personnel about various types of events.
IT staff could spend months, or even years, writing and testing homegrown integration scripts or pay integrators hundreds of thousands of dollars to get these systems to work together. On the other hand, they could implement a workload automation solution that can:
- Schedule and track jobs across multiple platforms
- Trigger processes for all of the company’s applications
- Monitor all workflows with a single, graphical interface
- Represent process control for all end-to-end business processes, including alerts that predict potential problems
The result: more reliable IT operations, faster integration, and lower IT costs.
Accelerating Business Processes
Today’s customers expect immediate service -- particularly when they transact business online. Organizations that want to compete successfully must become a real-time enterprise (RTE). Successful RTE environments rely on flexible infrastructure capable of supporting increasing demands in processing with minimal latency. This requires solutions that are capable of managing varied systems and applications from both IT and business perspectives.
To become an RTE, IT departments must create a real-time infrastructure (RTI). An RTI is a set of systems, applications, and management tools that can respond in real time to the business’s changing needs. In an ideal world with unlimited budgets, that would mean developing an entirely new IT infrastructure from the ground up, but in the real world, it means finding a way to get existing systems to work together seamlessly.
That’s where workload automation comes in. By using event-driven rules, workload automation can trigger work based on nearly any event -- including a Web services request, a resource usage level, or the arrival of a message. Instead of writing a complex program to trigger cross-platform processing, IT staff can define workflows through a GUI, simply and inexpensively linking processing events to each other across disparate platforms.
The self-healing and self-managing capabilities of workload automation solutions can also reduce the need for staff intervention. Such solutions can find and correct errors automatically while regulating workflows to meet SLAs. They can dynamically allocate resources to support business priorities. These capabilities help accelerate responsiveness to internal and external customers while reducing costs and risks.
Easing SOA Implementation
Workload automation can accelerate and simplify the implementation of service-oriented architecture (SOA) technology. Companies are moving towards SOA to readily leverage information and application functionality from anywhere. The engineering of a true SOA environment, however, may require the creation of thousands of lines of code so that systems can interact dynamically. This means that IT staff has to spend a tremendous amount of time writing, testing, and de-bugging code -- in addition to managing and modifying it -- since applications and systems inevitably change over time.
With workload automation, legacy applications can be incorporated into an SOA without costly re-coding. Any event in a given application can be used to trigger a process in another application. In fact, workload automation offers the added benefit of allocating processing resources on target servers based on pre-set business priorities -- so that new SOA-enabled processes don’t disrupt critical existing processes.
Given that SOA initiatives consume as much as 40 percent of the IT budget at some data centers, workload automation is an obvious choice. With workload automation, SOA benefits can be achieved quickly and inexpensively -- without disrupting the entire IT organization.
Workload automation also can ease mainframe ownership burdens by empowering distributed computing staff with little or no legacy platform experience to perform a variety of mainframe workload management tasks with minimal training.
Summary
IT managers should assess how they can more effectively leverage the current generation of highly sophisticated and adaptable workload automation solutions. The ability to link events and processes across applications and platforms can be extraordinarily useful in today’s heterogeneous enterprise environments -- especially as pressures grow to do more and to do it faster. By appropriately applying workload automation, IT organizations can meet these challenges quickly, precisely, and with lower cost.
- - -
Keith Woodward is director of product marketing at CA. Keith has spent more than 10 years in the enterprise software and IT industry and currently works with CA's Workload Automation solutions. You can reach Mr. Woodward at keith.woodward@ca.com.
Denise Kalm is director of product marketing at CA and has 30 years of experience in IT, including work in application programming, enterprise systems management, and performance management/capacity planning. You can reach Ms. Kalm at denise.kalm@ca.com
.