The IT world is ablaze with automation, as organisations strive for improved productivity and accuracy at a significantly lower cost. In line with this, a lot of companies are experimenting with the use of robotic automation tools, but adoption is still in its infancy.
And by robotic automation, we are not talking about the classic robot design that we all know and frequently see in factories, but advanced software-based robotics that can interact with existing web pages, infrastructure and systems that were never designed with robots in mind.
No robotic arms, just a virtual machine with access to system infrastructure. Simply put, software robots work around the process and complement it – they do not disrupt or break it.
What is robotic automation?
The robotic process automation (RPA) concept is well-known and is now frequently found in virtual platform management tools, running in the background of self-provisioning and quickly moving beyond the basics of deploying and scaling virtual infrastructure.
In short, the idea is bursting out of the “faster, quicker virtual machines/physical deployment” niche and taking over day-to-day administration operations.
The concept has historically been carried out in large companies which do a lot of data collection, extraction and manipulation – think typical mail-in forms, insurance forms and the like.
Its core use case is in doing jobs that are straightforward and repetitive, such as processing forms and importing them into a digital environment, before setting aside paper copies that cannot be imported because the document is mangled or the text is difficult to read. These jobs were typically essential, but low-skilled, and have been almost completely replaced by automation.
Exploring the use cases
This could be considered “old-school” usage, however, because robotic automation is now coming for the more mundane tasks carried out by IT administrators. One of those examples is disaster recovery infrastructure testing. Again, this is quite a repetitive duty, making it ideal for automation.
While it is a necessary evil, disaster recovery infrastructure testing is very time consuming, not to mention expensive and prone to errors. Automating the failover into an isolated environment is the easy bit. The testing component, however, requires people to set up, configure and test the disaster recovery instance. The people required to manage and implement the setup cost money. Automating the configuration, setup and testing of the application means expensive people are no longer required, but still need to be managed in the classic supervision sense.
That revelation means much bigger changes in the way tasks are tackled. Because automation essentially provides pre-programmed “free people hours”, with individuals continuously testing the disaster recovery capability of isolated environments, mimicking and – more importantly – managing and reporting the results as if a human were driving the test.
There is more to it than that, though. Robotic process automation works with existing infrastructure – and there is no requirement to retool, although that does not mean it should be discounted entirely. Unlike human input, RPA is also available for use 24/7 – the quicker, faster and more accurate selling points for the technology strike again.
Some companies have taken it as far as using automation to have a rolling test cycle for disaster recovery events, dramatically increasing the frequency and creating more trustworthy disaster recovery failover processes for when the day inevitably comes that something goes wrong.
It can also rapidly scale as required. There is no “training and upscaling period” – it’s just a case of deploying more software robots within the existing software configuration. As such, scaling work horizontally becomes a simple process.
But disaster recovery is just one possible use case. Technology is emerging to handle and manage tasks that aren’t a good fit for simple scripting. An example of this is disk expansion, which is quite a straightforward process to scale, but doing so to hundreds of machines introduces risk and complexity. Automation robotics can be used to modify the disks as needed, as well as carry out checks inside the virtual machine and monitor the hypervisor level for available disk space, which is not a trivial process to try to repeat.
The complexity of doing this by hand, with all the associated actions that need to be repeated precisely, makes it ideal for such things. Resizing 7,000 virtual machine system drives would be extremely error prone, not to mention expensive, if done by hand. Tasks like this can be made programmatically less difficult when done by a software robot.
Therefore, a lot of companies are using automation to permanently carry out functional testing on disaster recovery sites. Such functionality would not have been possible going back even just a few years.
The changes are coming to production infrastructure too.
An associated benefit of automation is the increased scope for analytics. Each process produces a wealth of data, such as process duration – for example, how long a failover took to complete – so a complete log output for further analysis can be compiled and bundled up to create effective reports. If humans were to do this, the data would be a lot less accurate, if it existed at all, because someone would need to keep note of the very many subtle timings.
We are just starting out on the journey to software robotics. Currently, we are only taking the initial steps to automation, and the systems have to be trained to perform a task. People may think about these as akin to a macro that we all know and love. The eventual destination is that the software will truly understand what it is doing.
How far off that is, no one can say for sure, but any administrator who wants to stay employed needs to ensure they are ahead of the curve.
CIOs, prepare for the sweeping changes robotic process automation will bring
Technology leaders driving innovation should keep in mind the wide-ranging impact automation technologies will have on their businesses. These include artificial intelligence-enhanced robotic process automation (RPA) and software defined infrastructure.
Automation starts deep inside a company’s infrastructure. In fact, the modern CIO’s infrastructure is now largely based on software. Infrastructure-as-code, including containers, has become the new foundation of modern enterprise infrastructures.
Artificial intelligence (AI) is only one dimension of automation, which also includes many sub-AI-level algorithms and tools. But increasingly, AI suffuses and boosts many different automation technologies – or will do so in the near future.
While RPA is creating a lot of operational value today, its future lies in cognitive-AI enhancement, a technology convergence that will solve many more business problems and will do so in a more sophisticated fashion.