The IT Process Automation Landscape
One of the most popular buzzwords today in IT management is process automation. You hear it all the time from vendors and we also hear it a great deal from our IT adopter clients in various contexts. Process automation, data center automation, run book automation, and sometimes workflow, often get mixed together in an unclear and ill-defined mix of possibilities.
Then if you add in other terms like job scheduling, automated configuration management, automated patch management, and even service request management, you start to get something beyond a soup with a curious mix of ingredients, and something more like a panoramic though surreal landscape.
In this column, Id like to examine some of what EMA is exploring (we, ourselves are not fully done with this discussion by any means) as the process automation landscapeboth in terms of how it exists today and where it may lead in the future. Along the way Id like to highlight some of the more tangible benefits and interdependencies.
First of all, Ill put a stake in the ground by saying process automation comes in three logical flavors: machine-to-machine interactions, human-to-machine and machine-to-human interactions, and human-to-human interactions. All three of these areas can provide dramatic benefits, but as automation becomes more pervasive, they may also lead to virtual train wrecks, as automated processes begin to collide. These areas of automation are also arising out of different domains and different technology markets. Probably the most prevalent two being data center automation (DCA) including network operations, and the service desk.
So, lets start by looking at DCA, which is at the heart of process automation. And the very heart of DCA is probably configuration management. This includes configuration management for servers, desktops, network devices, security devices, storage devices, OS configurations, application provisioning across the data center, as well as patch management updates as appropriate across the infrastructure.
It also includes the ability to automate support for compliance and security requirements including vulnerability assessments. A key part of this is automated checks for interdependencies and the automated reconciliation of configuration changes to validate that intended versus actual changes are aligned. If all of this sounds a lot like values potentially associated with configuration management database system (CMDB) investmentsthen youre already ahead of me. Were going to revisit that point at the end of the column.
And finally, from a configuration perspective, process automation will become critical for integrating virtual and physical environments, which by the way will be one of the areas of differentiation for CMDB technology providers in the coming years.
DCA also includes other technologies. Chief among these are job scheduling, and IT process orchestration, sometimes ill-named as run book automation. This last, in particular, is one of the foundational technologies that will help to span machine-to-machine, machine-to-human, and human-to-human requirements for automation as it matures. Another critical area of DCA automation and automation more broadly is chargebackalso ill named because it really is about understanding how your consumers value and utilize your services and how that impacts your capacity and cost requirements. In other words, generating a bill may be only an incidental concern. And finally there are areas such as database automation, capacity planning (an underserved frontier at present), and remote controlautomation via out-of-band technologies.
This is already a long list but were not done yet.
The service desk is the other large cluster of automation-related technologies. These include workflow, of course, where human-to-human process automation in support of best practices has its roots. Other areas of automation have to do with asset and inventory discovery and financial planning, as well as project and portfolio management much of which is similarly dedicated to cost optimization across IT.
But one of the most potentially game-changing technologies over time is associated with service catalogs. Today, most IT organizations are at least aware of the potential for service request management (SRM) to automate and enable self-provisioning between end users and IT for fairly basic, desktop-associated access to services. These may include cost and time delay information along with guidelines on usage for the end user. But service catalogs may also begin to model the business side of services from a financial and business impact perspective, informing not only end users, but also strategic service managers and line of business planners.
Automation, ITIL and the CMDB
They may also begin to include linkages to CMDB interdependencies that can help to automate the provisioning of far more complex services, say provisioning SAP to remote offices throughout the Baltic. This is probably two to five years away, but as service catalogs evolve to become something like the pretty face of the CMDB, the stage is set for heretofore unimagined levels of automation in the mid-term future.
In respect to ITILs lifecycle approach, EMA also recognizes two other general areas of process automation: lifecycle application development and service management. The former is evolving to support application developers with more automated tools and process for working to support real-world infrastructures in collaboration with operations and QA/Test. The latter, service management, is actually familiar ground, with capabilities such as automated discovery, application dependency mapping, automated cross-domain analysis and diagnostics, and automated impact analysis.
The benefits of automation are striking, even given the fact that most automation has heretofore been largely tactical and siloed in nature. According to our research, the top realized benefits of process automation are in ranked order are: reduction of human errors, improved security and reduced risks, cost reductions, reduction in operational complexity, improved compliance and security audits, and freeing up of strategic resources.
In other words, the dominant realized values are supportive of governance and reduced risk. This, I think, is telling in terms of the future of process automation and its immediate values today. Also of interest, is that EMA research shows that an integrated approach to process automation versus the more siloed, tactical approaches, demonstrate a three-times improvement in terms of Mean-Time-to-Repair and an improvement in availability of from 99.95 to 99.99. And while this data is directed mostly at systems specifically, the implications for a more service-oriented approach are significant.
And this leads me back to, as you may have already guessed, CMDB. Ultimately the future of strategic versus tactical process automation will depend on CMDB investments. As these become modularized, more flexible, and more extensible to respond to individual domain requirements as well as a centralized strategic vision, process automation and CMDB should become increasingly two sides of the same coin. Not only will CMDB provide the ultimate context for process automationto avoid those automated train wrecks, they, themselves will empower (as they are today) certain process automation capabilitiessuch as more comprehensive and effective automated audits, more effective automated configuration and patch changes, lifecycle asset management with increased levels of automation, and reduced MTTR by capturing asset-to-owner-to-service impact interdependencies.
And finally, CMDB are themselves dependent on process automation. EMA consulting engagements have already revealed that CMDB deployments without creative attention to automating repetitive tasks may prove very difficult to sustain.
If all this seems like a whirlwind tour I apologize but maybe thats in keeping with process automation itself. It is evolving both from the bottom up, through grass roots technologies supporting individual professionals, and from the top down, through strategic, service-oriented, cross-domain initiatives. As these two areas meaningfully begin to converge, tighten your seatbelts and hold on for the ride. The promise is huge, as long as its coupled with common sense and attention to process planning.
Dennis Drogseth is vice president of Boulder, Colo.-based Enterprise Management Associates (www.enterprisemanagement.com), an industry research firm focused on IT management. Dennis can reached at firstname.lastname@example.org