ITPA Defined (Sort Of)

Like Data Center Automation, the term IT Process Automation (ITPA) is often misused. So much so that in order to define it I will begin by telling you what it isn't.

It is not job scheduling, or workload automation, or whatever new term is being used to describe the category of tools historically used to automate batch processing. Trust me on this one, I spent a decade in the job scheduling business building and selling some of the most innovative and successful products in the market. These tools are valuable and essential parts of any IT automation strategy but they do not fulfill the needs for which dedicated ITPA tools are intended. Specifically although their automation processes look like workflows they are not. Job schedulers do not execute workflows, they manage dependencies between jobs. This is a subtle but critical distinction. Because they do not contain full workflow implementations they cannot represent complex logic (the kind that define IT processes) without layering drastic amounts of scripts and code on top. They also do not have the appropriate sorts of integrations (or adapters, or in job scheduling parlance agents). True enterprise class job schedulers do a great job at covering many operating systems from Windows to the mainframe and they have very deep integration into commercial ERP applications like SAP and Oracle but they do not integrate broadly or deeply with the IT tools in your datacenter such as event monitors, service desks, virtualization, and provisioning and configuration tools. Without these integrations you cannot automate your IT processes without layering scripts and code on top (are you beginning to see a pattern?). Finally job schedulers typically excuse themselves entirely from the task of handling the data passing between jobs in a flow or in ITPA terminology activities in a workflow. They may offer simple symbolic variable substitution but for the most part they assume that the data between jobs will be passed either in a database or an external file system. Any ITPA solution that does not offer native data handling that goes beyond simple variables will necessitate an inordinate amount of (you guessed it) scripting and coding in order to make it work. There are many other differences between job schedulers and ITPA solutions (including many things that job schedulers do better than ITPA like complex calendar handling and forecasting) and there are in fact many similiarities (especially from an enterprise architecture perspective), but these three major differences - workflow, integrations, and data handling - define the major differences between the two. These are not simply features to be added to job schedulers to transform them into ITPA solutions thereby finding a new market for mature products, they are genetic differences that define separate species. So if someone tries to sell you a job scheduler (under whatever banner) to automate your IT processes simply ask him or her how it handles complex workflow, integration with standard IT tools, and integrated data handling, without additional scripts or code. If they have a satisfactory answer then they will have cracked a code I thought uncrackable. If not, then you'll have saved yourself a lot of wasted time, effort, and perhaps money.