Cloud computing offers organizations the ability to dynamically scale IT infrastructure resources up and down on demand. However, those changes often need to take place faster than IT administrators can perform them, especially without enough coffee. As a result, the increase in companies relying on cloud computing, in turn, creates an increased reliance on IT automation technologies.
That can be problematic on multiple levels for many IT organizations. While any number of IT professionals take pride in their ability to write custom scripts that automate some portion of the IT management process, those scripts sometimes refuse to scale. Moreover, the scripts often can’t be easily replicated across the enterprise and—just to make things that much more interesting—are usually undocumented: when the IT pro who authored the code decides to leave the organization, the script disappears along with them.
A number of companies—including VMware, CA Technologies, Kaseya, Puppet Labs, UC4 Software, Advanced Systems Concepts and just about every server, storage and operating system vendor—have seized on this issue to one degree or another. But even those attempts to smooth out the automation process have sparked controversy.
Software vendors argue that server vendors’ automation solutions only involve the latest platforms, while their own approach to automation can apply to both new and legacy system infrastructure. Indeed, while server vendors such as IBM and Hewlett-Packard have pursued both hardware- and software-centric approaches to IT automation, there are questions about the degree to which their offerings will automate systems from rivals such as Dell.
The second major issue with automation is the level of disruption it brings to the IT organization. Today, thousands of administrators manually perform routine configuration and management chores that will soon be automated out of existence. As they do so, they neglect those IT tasks with more business value—tasks for which many of them have not yet been trained. The job for fulfilling those tasks could be assigned to new employees, potentially placing the administrators’ jobs at risk.
Obviously, that creates a scenario where individuals have a potential incentive to slow down the adoption of IT automation, if for no other reason than to buy some time for retraining.
Automating IT Management
But trained for what? IT automation fosters the convergence of job roles within the IT organization. People dedicated to servers, storage and networking tasks can now easily manage any and all of those functions, and more besides. Virtualization administrators can handle most hardware functions. And just to make matters more interesting, the whole “DevOps” debate is leading to the development of technologies that increasingly give developers more influence over the production environment. The demands of cloud computing are accelerating the convergence of all functions.
Cloud computing’s impact on administrator jobs could very well prove profound. A recent “New Skills for the New IT” report from Gartner predicts that, whereas 70 percent of IT resources are devoted to operating IT infrastructure today, by 2020 just 35 percent of resources will end up used in operations.
The cost of labor remains the single biggest cost of IT, an issue many organizations have tried to address by relying more on automation. But according to Dave Hitz, co-founder and executive vice president of NetApp, automation can’t proceed without a clear idea of how it will impact the organization as a whole. “If you don’t know what your IT processes are just yet,” he added, “you need to make sure that whatever automation tool you use make it easy to change your processes later on, just in case you’re wrong about what you decided to automate.”
In other words, before anything can be automated, the IT organization should take the time to actually define those IT processes. (In fact, when you peel back much of the conversation surrounding IT management initiatives such as the IT Infrastructure Library (ITIL), the primary goal of adopting such a framework is to better define IT workflow processes.) Whatever path chosen in the end, one thing is certain: after decades of using IT to automate almost every process on the planet, the time has come to use IT to automate the management of IT.