Today’s cloud was made possible by virtualization technology, which creates a software-based representation of hardware equipment. Virtual machines, such as those popularized by VMWare and the hypervisor technology that manages VM execution, make it possible to run different software on the same machine.
This concept is now expanding beyond the cloud to the physical world through the use of software that controls autonomous robots. I call this software-defined X: any physical task (X), from cleaning the floor at an airport terminal to delivering an item from one end of a warehouse to the other, can now be controlled through software. This is really taking “digital transformation” to its logical conclusion.
In the cloud world, virtualization conceals any physical aspects of a computing platform and instead presents users with a ”virtual” computing platform. So-called public cloud providers like Amazon Web Services, Microsoft Azure and Google Cloud Platform make it possible for developers to “spin up a VM” (i.e., create a virtual machine instance) in seconds rather than the days that it would take in the past to procure a new server computer, connect it to power and network, install software and carry out the myriad tasks required before it’s ready for use.
Beyond computing, this concept expanded to include software-defined networking (SDN) and software-defined storage (SDS). Software-defined networking uses software-based controllers or application programming interfaces (APIs) to direct traffic on the network and communicate with the underlying hardware infrastructure. Similarly, software-defined storage separates the management and provisioning of storage from the underlying physical hardware.
The combination of these technologies has enabled highly scalable, flexible and efficient cloud infrastructure, which is agnostic to the specific applications that run on it while adapting to their specific needs. While the actual functions of a social network, a streaming video site and customer relationship management software vary drastically, they can all run on this common infrastructure, which can adjust dynamically based on usage needs.
But all of these applications are just pushing bits around a screen. They can’t impact the real world. They can’t put food on our table or a roof above our heads. Until now.
A new generation of smart robots is emerging, and they come in all shapes, sizes and functions. It’s like a Cambrian explosion in robotics, with new robotic capabilities announced weekly by companies from Silicon Valley, Boston, Pittsburgh, Odense, Shenzhen and more. The use for robotics is also exploding across all industries, with thousands of companies building a robot for X.
X usually stands for highly targeted tasks, like in the case of robots that are designed and optimized to harvest specific crops like strawberries or to set up drywall in construction projects. In other cases, X may represent a more generic task, like transporting items, which can be applied to very different scenarios ranging from moving linens at a hospital to carrying items within a warehouse to get them ready to be shipped.
In order to manage this advanced equipment, a new type of software platform is emerging that abstracts the complexity of a specific robot (the hardware) while making it possible to control its operation. Large enterprises can use this software layer, or “middleware,” that sits between their internal systems and the robots, to introduce automation into their existing physical processes without having to hire armies of roboticists, the sought-after engineers who specialize in robotics.
This approach has many benefits, allowing enterprises to simultaneously drive productivity and flexibility. For instance, applying the principles of infrastructure as code, it’s possible to define complex behaviors and outcomes in software code, which is then translated into lower-level commands that are ultimately carried out by robots. This allows repeatability, auditing and optimization, which are critical to driving continuous improvement.
Picture an autonomous floor scrubber, which is like an oversized Roomba and increasingly being used to clean airports, big-box retailers and warehouses. Instead of having a member of the custodial staff press a button to get it started, cloud software can use data from IoT sensors to determine where the robot should clean and how often, sending the commands to the robot. Moreover, the software can monitor which areas have been cleaned down to a squared inch, keeping a record and optimizing the routes and frequency over time.
A growing ecosystem of tools ranging from observability and orchestration of robots to integration of data sources from the internet of things is developing to help companies pick just the right pieces they require to address their specific needs. Beyond the actual programming and tools, a set of best practices is also emerging. For instance, the Robot Operations Group is a community of expert practitioners that is creating open-source recommendations, such as their Robot Operations Manifesto.
This is not science fiction. Leading companies are already starting to deploy automation at scale, and so can you. Start by looking at your operations to identify the tasks that, if optimized, could have the biggest impact. Some of these will be common across industries, like moving material more efficiently, collecting data about the world or carrying out repetitive or dangerous tasks. There’s probably a robot for that.
But the hardware is only part of the solution. You need a software platform that will scale with your needs, that will orchestrate heterogeneous robot fleets, that will handle massive amounts of data in real time and that will integrate with your existing enterprise systems, from ERPs to building management systems.
The evolution of software is coming full circle. As the modern internet was going through a similar explosion 25 years ago, Nicholas Negroponte, co-founder of the MIT Media Lab, published the influential and prescient book Being Digital. His main thesis was that we should make the shift from atoms to bits, elevating digital content at a time when news, music, movies and books were physical objects. Now, for the next big shift, we can use bits (software) to move atoms (stuff) through autonomous robots. Welcome to software-defined X.