Sergey Nivens - Fotolia

IT automation changes the game, but limits remain

Automation tools give admins ways to automate routine or even complex tasks, but some limitations mean IT must keep their hands on the wheel. Heuristic automation will take over areas where traditional deterministic rule sets cannot keep up.

ORLANDO -- IT automation tools have become so pervasive and important that a growing number of businesses employ IT staff just to manage automation.

These tools touch servers, IT processes and workflows, application releases cycles, cloud resources, mobile devices, networks, scripting and even business processes, according to Ronni Colville, analyst with Gartner Inc., who spoke about automation at Gartner's IT Operations Strategies and Solutions Summit 2015 here this week.

Colville explained the factors that drive this push to automate, what automation issues still haunt IT organizations and how these tools continue to evolve.

Efficiency and cost lead the charge

Colville's research reveals efficiency is the single biggest driver behind IT automation adoption. IT staff need to perform an ever-growing array of complex tasks that involve users, local and cloud-based compute and storage, applications and so on. IT automation tools can leverage established rule sets and available instrumentation that allow administrators to handle those tasks quickly.

For example, workflow automation tools can ensure all necessary approvals and licenses are present, provision a VM that provides the necessary resources and then send follow up messages that the work has been performed. Consequently, automation allows a much higher degree of scalability in IT processes and operations than would otherwise be possible with manual processes, allowing smaller numbers of IT staff to manage much larger infrastructures.

Cost reduction emerges as another important reason to adopt IT automation tools, but Colville notes it's not the tool itself that saves money.

"Cost reduction is about making an application more efficiently delivered," she said.

For example, getting patches out to every constituent system in a consistent and timely manner can prevent costly errors that might be caused by manual patching processes -- leading to lost production and lost business.

Risk mitigation is a third prominent driver for IT automation, but the underlying issue of risk in this context is really about improving supportability and minimizing potential disruption of services -- two common problems that occur when IT tasks are handled in a manual ad-hoc approach.

"There's less one-off tasks to handle," Colville says.

For instance, an IT administrator can easily provision a LUN using a command line, but errors or oversights in the command line can be difficult and time-consuming to troubleshoot. By comparison, even a PowerShell script allows complex combinations of commands and logic to be reviewed and documented by the entire IT staff. There are fewer errors that would need to be fixed.

There are additional factors to consider in automation tool adoption, such as predictability. This ensures the same process is invoked the same way every time, and administrators have a set expectation of the outcome.

A close corollary here is using tools to establish best practices for IT and the business. It's not just doing things the same way, but doing the "right" things in the same way. Since tool activities are logged, it is possible to trace actions to individual administrators and bring a greater level of accountability to the IT environment. You don't just make a change, you know who made the change and when. Taken together, predictability, accountability and adherence to established best practices can be essential elements of a business' compliance strategy.

IT automation challenges remain

Although the motivations behind IT automation tools are clear, Colville explains there are also numerous issues that still conspire to limit adoption and productivity.

The most notable problem is a shortage of people and skills. IT teams are already overworked, so the know-how and time available to deploy and configure an automation tool may pose an insurmountable obstacle for some organizations. In other cases, cultural resistance may impair adoption as administrators hesitate to embrace tools they fear might ultimately render their jobs moot. This is not the case, since the efficiency gained with automation tools can free administrators' talents for more strategic and beneficial projects within the business.

Another collective problem is that the IT staff simply doesn't know how to implement an automation tool. Colville points to a lack of existing process documentation, so even though processes are performed a certain way, they aren't described in any comprehensive or complete manner. This makes it extremely difficult to translate processes into automation logic.

A closely related issue is the lack of process and workflow creation expertise where the IT staff might see the process written out, but doesn't have enough knowledge of the tool and its capabilities to translate the manual process into the rule sets required to accommodate the particular tool.

A great many more organizations face immature infrastructures or operations. This usually occurs because the business is changing or growing at a rapid pace, and leads to documentation and skill set problems. Smaller organizations may simply be unable to make a strong business case for automation tools.

We have more tools than we know what to do with, but picking the right ones and getting them to work together with their dependencies is a problem.
Director of Operations for a global finance company

Finally, organizations must frequently grapple with multiple tools from a variety of vendors. Sometimes multiple tools are acquired as a bundle, but they may also be acquired through mergers and acquisitions or purchases through other departments. IT staff may find themselves in situations where multiple tools are available to accomplish the same tasks, and it can become extremely difficult to consolidate duplicate tools without time-consuming retooling.

When multiple tools must work together, Colville says organizations still face a lack of standards and standardization that can prevent otherwise complementary tools from working properly together -- further complicating the automation landscape for a business.

"We have more tools than we know what to do with, but picking the right ones and getting them to work together with their dependencies is a problem," said one director of operations for a global financing company.

Heuristics is coming

IT automation tools are still evolving and traditional deterministic rule sets are quickly giving way to greater autonomy and intelligence. The human knowledge present in traditional rules is increasingly merging with other data sources like unstructured text, machine learning, and enterprise instrumentation to provide heuristic automation tools that can monitor and learn in response to dynamic changes in the environment.

For example, workload automation tools will move beyond preset thresholds for workload balancing by monitoring conditions through logs and real-world sensors, making its own statistical determinations of "normal" conditions, and only when those dynamic settings are crossed will the tool take action. But ultimately, Colville says that any automation initiative must serve the business.

"You must change the metrics to measure how automation benefits the business; not IT," she said.

Stephen J. Bigelow is a senior technology editor in the Data Center and Virtualization media group at TechTarget.

Next Steps

What is the new IT automation imperative?

IT automation tools to the cloud

The right time to consider IT automation

Dig Deeper on Data Center jobs and staffing and professional development