In the beginning, IT ran on one platform—the mainframe. Workloads were predictable and stable, and we could schedule accordingly. But then came UNIX, AS/400, Windows, and many other platforms, all with accompanying workloads. The Internet and the need for non-stop availability changed the workload model; “stable and predictable” flew out the window. Yet the data center kept the same goal: to manage the workloads to provide the highest level of service at the lowest cost.
The best way to provide optimal service at a reasonable cost is to automate, but automating the data center is easier said than done. Over the years, you’ve adopted a variety of platforms, applications, tools, and processes that may have worked well independently but don’t necessarily work efficiently together. Ever-changing workloads present challenges we couldn’t have imagined a few years ago. You may realize the tools and processes you have are costly in terms of overhead, and it’s hard for your organization to grow and remain competitive when you’re burdened with extra costs.
Adding more people to handle more work isn’t always the answer; eventually, you reach a point of diminishing returns. At what point does automation become less expensive than current tools, processes, and people?
Dynamic Workload Management
Workload automation has evolved from traditional job scheduling based on date/time criteria and static business needs to a dynamic model based on fluctuating workloads and changing business needs. Effective workload automation coordinates varied workload types with complex dependencies across multiple operating systems and platforms—in real-time. While date/time dependencies still drive some workloads, real-time IT processing brings a new set of workload challenges. Some IT organizations have been trying to meet these challenges by adding more people, but if you add people without also implementing intelligent automation, you can quickly reach the point of diminishing returns.
To automate cost-effectively, you need a single workload automation scheme and framework that meets the needs of your applications, your multi-tiered infrastructure, and your business. Effective dynamic workload management:
- Supports all platforms and applications in your enterprise
- Provides a consistent look and feel for managing platforms and jobs
- Shows how workloads affect your business goals and Service Level Agreements (SLAs)
- Enables planning and provisioning.
Let’s look at each of these areas more closely.
Support all platforms: How many platforms does your IT organization support today? How many applications touch multiple platforms? What integration points do you have? How much have all these changed in the past year? You need a solution that provides enterprisewide, end-to-end workload management.
The different platforms aren’t the only issue; processes often differ between platforms. Enterprise Resource Planning (ERP) and custom applications add more complexity. The nature of batch production itself has changed from a date/time model to an event and transaction model. Standardizing and automating across platforms, applications, and the batch jobs being processed can be a gargantuan task. Adding to these complications, mergers and acquisitions require IT to quickly absorb or integrate varied, geographically dispersed infrastructures.
An effective, dynamic workload automation solution integrates management of critical workload processes from a single point of control. Cross-application and cross-platform scheduling capabilities (such as job dependencies, workload balancing, and event-based job execution) enable business growth and prevent scheduling problems from developing into business problems. An effective dynamic workload management solution eliminates the need for multiple schedulers, providing a higher level of integration and a lower learning curve. It enables integration of applications, including legacy, ERP, Customer Relationship Management (CRM), Java, Web services, and more.