Transforming Your Workload Automation Capabilities

As the new Transformers movie hit the theatres this past summer, we were reminded about the "transformation" we have seen take place over the last twenty-plus years with regards to job scheduling and data center automation. Job scheduling has truly transformed itself from a mediocre looking auto"mobile", into a powerful, conquering, efficient, workload automation machine.

What began as a simple task-oriented system to "batch" workload processes together and initiate them at a date/time when the online systems were down and the processing could take place, has now become a tool that must allow for processing at any time, based on random business transactions done over an internet-enabled device from anywhere in the world.

It's really funny to see the various stages of workload transformation that companies are in today. The Next Generation Data Center are all about processing customer data immediately and eliminating the restrictions and limitations still in place in some long-term IT shops. Some of these restrictions are necessary and a result of not being able to take the business risk replacing the outdated systems. But, some of them are simply still in place, because "that's the way we've always done it". When talking to a data center manager a few months back, I asked him if they ever thought of processing a randomly arriving transaction as soon as they received it. He was taken aback, "Oh No, we would never process any unexpected file that arrived", he told me. As our discussion about "business-initiated" workload automation continued, and I shared with him the advances in the technology that were enabled largely from the product architecture, he looked at me and said, "You know what, I bet our customers would love it if we could process their transactions immediately!" Sometimes all you need is a little insight, an open mind, and transformation can occur.

Then other times, you might need a technology change as well! Many of the more widely installed workload automation solutions are based on the initial method of "batch processing", and some have only recently added the ability for ftp file and dataset monitoring at creation time. Still others have found themselves "retro-fitting" more immediate triggering mechanisms to their batch scheduling products, and charging you for the additional components they create. But, few have really been able to tap into the reality of "business-events" initiating workload processing at very granular trigger conditions. Even then, conditions must be set and in many cases VSAM files must be scanned to determine if any activity is ready to be processed. The overhead of this activity on a system that was not designed for this could be cost-prohibitive from a simple resource perspective. You have to think about the cost of those resources being consumed. So, what is a data center to do for their new business, which is really driving their company growth and profits, while still supporting the "legacy" workload processing which is more batch-driven? Does it really have to be a total 'rip and replace' of one system for another, in a 'big bang' conversion? Not necessarily. It definitely reduces the risk by doing a big bang conversion, but without 'reengineering' the workload to take advantage of the newer automation benefits, you are not gaining the most value from the new technology. In short, you are not getting what you paid for. Yeah, I know, some companies want to minimize the risk and opt to push the "reengineering" phase off to a later date (that NEVER happens, by the way). I would advise companies to think of the process as "reenergizing" versus re-engineering, and it actually serves to provide dual benefits of a better processing method for the business data, but it also better educates and helps your production control or automation personnel to embrace the newer methods, rather than them trying to get the newer product to allow them to work the same way as the product you are replacing!! I have seen this scenario play out many times, and while the "big bang" converted customers get over to the new product quickly, in the specified timeframe, they are often left feeling remorse because they never get to optimize the workload implementation to take advantage of the products capabilities. As new releases with even newer functionality are delivered, they just find themselves getting further and further behind.

Transformation is a process, and doesn't happen overnight, or as quickly as shown in the movies either. But we often heard that "anything that is worthwhile doing, is worth taking the time to do right"! Updating your mission-critical workload processing to business-initiated methods could be vital to remaining competitive in your market, providing better service to your customers, and increasing shareholder value for your company. Start with the workload that needs the transformation the most, and work your way thru the process, you will end up with a much better solution,and a more knowledgeable and more automation-focused staff, once the journey is complete. Having taken both paths, converting big bang, and transforming by re-energizing and taking advantage of the tools capabilities, I would choose the latter, every time. You simply leave too many of the benefits of converting in the first place, lying on the table.

Transformers, now that is a movie I would see again!

Entry Permalink

Tweet