informa
/
AI/Machine Learning
Commentary

How IT Organizations Are Using Automation

IT organizations already have a lot of automation in place. Here's how they are updating their existing automations and blazing new paths.

The number of organizations deploying automation at scale has tripled in the space of two years, according to a December 2020 Global Intelligence survey by Deloitte of 441 executives from 29 countries. The survey further found that COVID-19 accelerated the need for automation solutions -- particularly those that offer options for scalability and rapid deployment -- within enterprise organizations.

For CIOs responsible for implementing IT automation now and in the future, deploying solutions that are already proven, and then identifying “next step” automation that can better IT performance, will be central to improving operational agility and execution.

Here is what we’ve learned works well, and this is where IT automation is headed.

Today’s IT Automation Successes

In many ways the processes that IT has chosen to automate over the past two years already were semi-automated, but they had room for greater automation improvement. In other cases, newer technologies such as IoT (Internet of Things) created additional opportunities for automation that weren’t widely available even a few years ago.

The following are automation use cases that have been deployed over the past two years, with operational payoffs for IT:

Batch Process Automation

Those who grew up in the mainframe era will remember JCL (job control language) as the “glue” that pieced together nightly batch run processes. JCL continues to run on mainframes in enterprise shops today, although many IT departments have taken the time to streamline their JCL for optimal performance in a lights-out data center.

For a majority of companies, however, batch process automation is now known as workload automation. There are many different choices for workload automation that not only run nightly batch jobs, but also the many processes that must be performed within IT infrastructure during daytime hours.

In the real-time, daytime environment, automated workload management automates many behind-the-scenes infrastructure processes that IT no longer has to oversee or perform manually. For instance, IoT data can be streaming into the enterprise with alerts and queries coming from it, but at the same time, there might be other data in the stream that the company wants to use later. The automated workload software can save this excess data to disk so it can be used later, with no need for IT to intervene.

Automated System Provisioning for Application Testing

A historical bottleneck for application testing has been the wait time developers had to endure while data analysts and system programmers prepared test environments that the developers had requested for their apps. Depending on these individuals’ schedules, an application developer could wait several hours or even several days before a test system was provisioned. This delayed app time to market and frustrated developers.

Corporate IT has overcome this bottleneck by moving to automated provisioning of virtual test environments. This has eliminated most of the time requested from data analysts and system programmers and has sped application time to market.

Automated QA Testing

Corporate IT has adopted automated application testing software for key steps in the application QA test process. This includes unit testing, integration testing, and regression testing. By using automated QA tools, QA departments reduce the amount of time required for manual testing. Much of the manual testing is replaced by a set of test scripts that QA designs for the different execution scenarios that an application has to perform and that the automation software executes.

Automated regression testing is especially useful. It test runs a new application in the target IT infrastructure. This ensures that the application introduces nothing new that would disrupt or nullify the functions of the infrastructure data, routines, and resources in the new environment.

Given the continuous revised and new development coming out of the DevOps methodology that is now common in IT, ensuring that new and revised apps continue to work with existing IT infrastructure is paramount. Being able to automate regression and other testing time can save the QA test team between 8-16 hours per application.

What new areas of IT automation focus will we see over the next few years?

A More Holistic Approach to Infrastructure Automation

If the past two years have been spent on incorporating IT automation into everyday IT, the next few years will be dedicated to a more strategic and holistic approach to infrastructure automation that ensures that all automation tools work well together on a common platform. The end goal will be to automate repetitive Infrastructure functions with self-executing scripts that can reduce human effort. These scripts will cover both overnight and intra-day functions. There are infrastructure tools and platforms that work in the cloud, and platforms that work in hybrid on premises/cloud environments.

AIOPs Observability

The last few years have seen IT install a plethora of network and system monitoring tools that automate tracking and immediately issue alerts if irregularities are detected (e.g., an unauthorized network access). This is good, but it still requires IT to research and troubleshoot to correct the issue.

The next generation of these products will use AI to not only detect irregularities and issue alerts, but to also deliver a contextual understanding of how and why the alert occurred. As an example, a network access may be picked up from a West Coast office on a weekend, with the system also telling you the office is typically closed and off limits at that time; or a server or router on your network may be at near capacity, but the software also tells you that this asset is functioning at its peak time of day and that traffic will soon quiet down.

These added insights into the particulars of IT infrastructure situations don't relieve IT of all incident mitigation work, but they speed time to results because they’ve already performed some of the preliminary research IT would normally have to do.

Low- and No-Code Applications

Low-code development can save businesses hundreds of development hours, cut their development times by 50% to 90% compared to developing an application from scratch with custom coding, and reduce their IT spends by as much as 20%, according to Matt Duench, senior director of Product Marketing at Auth0, a security authentication and authorization platform.

No-code software abstracts the entire development process away from the need-to-know underlying IT infrastructure. It is especially useful for user developers because they can quickly write an app without having to think about underlying IT infrastructure needs, since the no-code software auto-generates all of the infrastructure code for them. Low-code software also automates much of the IT infrastructure coding for IT developers, which saves them time.

Both low- and no-code solutions owe their heritage to early 3GL and 4GL report generators, which began to appear 50 years ago and auto-generated COBOL code. Today’s low- and no-code tools go far beyond this. They can address multiple IT cloud and on premises environments and can generate code in languages that include Java, PHP, and Python.

Related Content:

How IT Leaders Can Adopt an Automation Mindset

Automation Revs Pandemic IT Toolbox

IT Automation: Still Room for Improvement in Some Spots

Editor's Choice
Cynthia Harvey, Freelance Journalist, InformationWeek
John Edwards, Technology Journalist & Author