One simple way to look at automation is that logically it falls into three classes: machine-to-machine, machine-to-human, and human-to-human. And of course some capabilities and products cut across all three. But there are fundamentally different traditions on the two extremes.
EMA, like many others in the industry, sometimes struggles to adopt a common nomenclature for slippery phenomena that cross silos - in our case that cross practice areas. Rather than seeing this as a drawback, though, I believe that just having the debates is positive and healthy.
Of course, we could avoid them and confine ourselves to only what’s obvious. But then, we wouldn’t really be doing our jobs - and some of the major waves transforming the management marketplace could wash over us unseen until we suddenly realize that we’re drowning (figuratively speaking at least) in an ocean of real confusion.
One such wave – one that’s already cresting, so get out your life rafts – is automation. There are plenty of free-floating terms/ideas/technologies out there on the automation tsunami already. For instance, there’s “Data Center Automation,” “IT Process Automation,” “Runbook Automation,” “Service Automation,” and one that EMA created primarily for internal use, “ITSM Automation,” to reflect that cluster of technologies that have grown up from service desks, service catalogs and workflow – and one might argue the CMDB as well.
EMA also has a term called “NOC Automation” which is simply a placeholder to capture any automation relevant to network operations - such as WAN optimization or dynamic routing. Other backbone technologies have been around for a long time - like configuration management, automated diagnostics and discovery - and in the data center, load balancing and job scheduling or workload automation.
You might ask, then, how do all of these (and other) pieces fit together? One simple way to look at automation is that logically it falls into three classes: machine-to-machine, machine-to-human, and human-to-human. And of course some capabilities and products cut across all three. But there are fundamentally different traditions on the two extremes.
If you think of human beings as “top” and machines as “bottom” then let’s start from the top-down. Service-desk workflows have been around for years, with trouble-ticketing solutions being the most obvious. But with the rise of ITIL best practices, the service desk (we sometimes hear the term "consolidated service desk"), is viewed as a center for automating ITIL-defined processes such as configuration, change and release management, or incident and problem management.
But this “tradition” - because it truly comes from a certain set of cultural roots - is very human-to-human centric. It’s focused on codifying how people work with each other and bringing in human-to-machine automation (configuration and release management, discovery, etc.) from this top-down view.
Service catalogs are worth mentioning here because they’re something of their own phenomenon. They hold the potential to become one of the centers bringing all these pieces together by modeling relationships – much like what is done in effective CI class and attribute modeling in the CMDB. Most visible today is customer-facing automation through service catalogs, including self-provisioning, or on-boarding new employees.
Here the catalog may contain time/cost and operational processes in its modeling that serve to both communicate service options to customers and streamline processes for fulfilling the services on the back end. As linkages between service catalogs and CMDB systems become stronger, this capability to automate more complex provisioning tasks (as an extreme example – provisioning an SAP application to all branch offices across the Baltic) will move several steps closer to reality and out of the realm of science fiction.
Now let’s look at the other dominant tradition, which I’ll call “Data Center Automation” for now. This tradition is not at all ITIL-centric and has stronger operational and machine-to-machine and machine-to-human roots. Job scheduling, patch management, database automation, and dynamic capacity management are classic examples here. Automated diagnostics - analytics to determine root cause - may also be aligned with DCA. And of course configuration/release management is a mainstay as well, but with more of handshake the ITSM area.
Here, possibly the most game-changing technology is sometimes referred to as IT Process Automation (ITPA), which is a classier and more appropriate label than “run-book automation.” Like service catalog modeling integrated with CMDB systems, it has the potential to bridge process automation from granular machine-to-machine interaction with top-down, ITIL-driven best practices over time. But for now most actual deployments are clearly more in the operational arena, with strong granularity in automating machine-to-machine capabilities.
By the way, this list is not complete. For example, EMA is also looking at automation technologies directed at servicing application developers/Q/A Test and production-level requirements to provide better, more automated insights into application design in real-world networks.
The net of it is that automation is not only multi-faceted and complex, it is an area of huge innovation and one that’s becoming more and more urgently relevant as virtualized infrastructures get paired with virtualized application designs – and SOA-based Web services all become network versus just data center realities. The tide has turned against naysayers based on EMA research – and many of you seem to be seeking effective levels of automation more actively than ever.
But I suppose you would know that this wasn’t a column from me if I didn’t end by stressing the importance of CMDB Systems. Without that cohesiveness, you may be automating in blind silos – or in other words you’ll be automating train wrecks. It was striking to me then, that based on June data across 174 respondents -- the single most sought-after attribute in CMDB technology, even ahead of discovery, was “links to process and workflow automation.”