Robotic Process Automation and the Future of the Smart Data Center

Data centers need to smarten up. Robotic Process Automation (RPA) will bring data centers into the future of automation through the power of robots leveraging AI and ML.

Brent Whitfield -
Illustration: © IoT For All

When many people think about a robot, they imagine a metallic humanoid or maybe one of those smart machines that compete against each other in TV shows.

When we talk about Robotic Process Automation (RPA), we’re just as likely to be referring to “software robots,” programmable or self-learning systems that carry out operations on other systems.

This article looks at the role of RPA in the data center of the future, the so-called “smart data center.” We explore why data centers need to smarten up, what types of RPA are available, how RPA can be integrated into the data center and the benefits of doing so.

Why Do We Need a Smart Data Center?

So, why do we need a smart data center? To put it another way, what’s wrong with existing data center technology?

The answer lies in enterprise IT trends. Businesses form the customer base of the majority of data centers, whether that’s a small regional data center, a busy colocation or the globally distributed network of huge data centers that underlie the public cloud providers.

As businesses wake up to the power and efficiency of the cloud, they’re setting up DevOps teams and microservices which demand real-time processing, elastic scalability, big data storage capacity and 99.99 percent or more reliability. To meet the new demands required by these business models while keeping costs competitive, data centers have to reduce overheads while improving reliability and performance. A tough ask!

As infrastructure becomes more complex and distributed, there’s an additional argument for robotic assistance. Humans are simply unable to monitor and process the multiple streams of information coming into a data center without making errors or reducing speed and performance. Network downtime is serious enough, but with data breaches now attracting record fines, mistakes can threaten the very existence of a data center.

As the Los Angeles IT consulting firm DCG Inc. put it, “Considering the amount of data that will be coming into and out of your existing network, every last bit of that data must be accounted for.”

With human error still the most common cause of network issues, robots offer a safe pair of hands, at least as far as the following complex, repetitive processes are concerned.

But, aren’t robots too dependent on their human overlords to adapt to changes in system states? To answer that question, we need to look more closely at how RPA works and the different types of software robots that exist.

Two Types of RPA

RPA is a broad field but generally consists of a controller, a set of developer tools and a software robot. Robots can be further broken down into two main types: programmable and intelligent robots.

RPA with programmable robots is an intensive process as every step of a process has to be broken down and the robot programmed to carry out the steps according to fixed parameters. This is adequate for simple, repetitive tasks, and this type of RPA is commonly found in the world of manufacturing.

RPA with intelligent robots makes use of machine learning (ML) technology. These robots monitor processes as they’re being performed and analyze vast amounts of historical and current data to learn how to optimize them. They can then be configured to carry out a range of actions based on high-level policies and awareness of the state of a system. They can decide which tasks should be performed and when they should perform them.

This latter type of software robot is increasingly required for the complex environment of the modern data center with its rapidly shifting parameters and vast amounts of unstructured data.

RPA technology is maturing rapidly to the point where it can now be integrated into a data center with minimal disruption, producing reliable results and speeding up mission-critical processes.

However, if you’re looking into RPA for your data center, you don’t have to rush the process. In fact, a piecemeal approach is often better because, as with most new technologies, the benefits may not be fully utilized without a thoughtful, methodical approach.

Moving From Manual Control to Full Automation

Fully automating a data center is still a future goal with many people still skeptical that robots can safely handle the many processes involved. As with automated cars, there will need to be plenty of testing and intermediate stages before an autonomous data center can be left to its own devices.

Integrating RPA into the data center can be thought of as a four-stage process from manually assisted operations, which already occur in most data centers, to full automation:

Assisted Manual Operation

This is where bots are used to automate simple tasks, reducing the number of steps a human operator needs to execute. The human operator will still need to trigger the initial action, such as to perform a backup or move data from near online to online systems, but the robot will follow scripts to carry out the sub-steps involved in executing the task.

Partial Automation

With partial automation, RPA and ML are combined to analyze trends and recommend actions to human operators. An example is Dynamic Resource Scheduling (DRS), where robots can recommend shifting workloads based on load balancing rules.

Conditional Automation

With conditional automation, robots can monitor the state of a system and then take over when certain conditions have been met. These can be specifically defined conditions (e.g. a certain time of day) or conditions that fall outside of a parameter. For example, if a robot detects temperatures in a part of the data center falling outside of a defined range, it could execute a range of actions to reduce the risk of damage, take remedial action and alert the data center controllers.

Full Automation

The fully automated smart data center of the future will use ML to self-learn and plan ahead using current and historic data and modeling scenarios. It will evolve over time, adjusting parameters based on experience and improving its capacity management, resource allocation and energy efficiency. The data center robots of the future will also be able to proactively monitor the health of the data center and take remedial action such as ordering, replacing and configuring hardware components.

The Benefits of the Autonomous Data Center

Robotics has already had a positive impact in the manufacturing sector. As DCG IT services Los Angeles explains, “Manufacturers are enjoying lower overall costs, greater precision, increased safety in hazardous operations and better productivity.”

Its integration into the data center has already begun and is likely to accelerate with cloud providers such as Google buying up robotics companies and working with RPA specialists such as UiPath to develop a range of automation options.

Benefits That Data Centers Can Expect From Embracing RPA Include:

  • Increased productivity (so-called ‘cobots’ are six times more productive than humans in carrying out repetitive tasks.)
  • Increased human value as employees are freed up to take on creative and strategic tasks (e.g. DevOps, business planning)
  • Lower risk of human error
  • Better capacity management
  • Better resource planning
  • Increased energy efficiency
  • Proactive maintenance
  • 24/7/365 monitoring and automated repair (self-healing)

The robots are gradually gaining a foothold in the data center, and it’s only a matter of time before RPA is sophisticated enough to bring us to the verge of the automated data center.

The potential performance gains and cost savings for data center owners and the businesses that use them are a cause for much optimism.

Author
Brent Whitfield - CEO, DCG Technical Solutions Inc.

Contributors
Guest Writer
Guest Writer
Guest writers are IoT experts and enthusiasts interested in sharing their insights with the IoT industry through IoT For All.
Guest writers are IoT experts and enthusiasts interested in sharing their insights with the IoT industry through IoT For All.