This guest blog post is written by Mike Fahrion, CTO, IIoT Solutions, Advantech. It originally appeared at https://blog.advantech.com/sites/iiot-us/extending-virtualization-and-cloud-native-architectures-to-the-networks-edge/
By 300 BC, the ancient Phoenicians had established a logistics model for international trade that changed very little for more than 23 centuries. Barrels and burlap sacks of goods were loaded onto donkey-drawn carts, then onto hand-crafted ships to be traded throughout the Mediterranean, the Black Sea and the coastal Atlantic Ocean.
The more modernized version of the Phoenician’s logistics model, termed “break bulk cargo,” required massive quantities of dockworkers. Ships and trucks spent more time loading in ports than in transit. Theft, breakage and lost cargo were rampant. In the early 20th century, it was often quipped that, “dock workers made $20 a day plus all the Scotch they could carry.”
In the post-depression 1900s, a trucking company owner recognized the inefficiency in his trade and began devising a standardized method for delivery of goods that would forever change international trade and globalization. In 1956, Malcom McClean sold his trucking company, bought and retrofitted a ship to carry his patented shipping container. Recognizing that standardization and ecosystems were critical to success, McClean issued a royalty free license of his container patent to the ISO organization and undertook convincing port authorities, roadway and railway shipping companies to adopt and support the new standard.
McClean’s innovation didn’t just increase shipping efficiency. Fully enclosed containers were safe and secure, eradicating theft and damaged goods. Containers streamlined the connection and relationship between the manufacturer and the consumer. Manufactured goods were sealed into containers at the manufacturer’s own dock, not to be touched again by human hands until the container was delivered to the distributor or customer over any combination of road, sea and rail.
Less than 20 years later, as McClean’s innovation was rapidly revolutionizing world trade, the seeds of another innovation were sown. Software partitioning technology was developed on early mainframe computers in order to meet requirements for support of concurrent users and applications. This was the first step in the journey toward software virtualization.
Clicking forward just two more decades and the innovation of virtual machines (VM) on open architecture x86 systems was a critical milestone in the world of server rooms and data centers, enabling the advance of scalable cloud-based computing.
Virtualization is key to scalability. It decouples software development and delivery from not only hardware, but also eliminates complex dependencies and conflicts with adjacent software operating in the same hardware environment. This decoupling has enabled enormous gains in the efficiency and speed of deployment in server level environments.
Yet, in today’s world of the Industrial Internet of Things (IIoT) and digital transformation, there is a rapid, shifting focus toward “the edge.” The edge is often a more constrained operating environment than the on-premise or cloud-based data center.
When we attempt to bring the value of virtual machines to this more constrained edge, termed the “smart device edge” by the Linux Foundation Edge (LF Edge) umbrella project, we expose a weakness. Each virtual machine requires its own operating system. This overhead quickly becomes burdensome once we leave the controlled, protected environment of the data center and move toward the edge, where we typically find constraints around size, cost, environmental conditions and ruggedness.
Another form of software virtualization, called “containers,” has been experiencing exponential growth by bringing virtualization to more constrained devices found at the “smart device edge.” The smart device edge is where ever-growing quantities of compute devices are found. These devices are deployed at the front lines where the physical world and the digital world intersect in the IIoT.
This is critical to the world of digital transformation and the IIoT, where use cases for edge computing are growing rapidly, yet the challenges of deploying and managing massive quantities of edge compute devices are only beginning to be realized.
Containers are a lighter weight version of virtualization. Unlike virtual machines, which require one operating system per VM, containers share a common host operating system. Each container holds only the application itself along with its dependencies and configurations. This creates lightweight application packages that are easy to deploy, easy to maintain and portable across hardware architectures. Lightweight container technology enables developers to extend the concepts of cloud-native architectures all the way to the smart device edge.
The concept of hardware decoupling, along with decoupled and loosely coupled applications is central to cloud-native architectures. The ability to extend those architectures to the smart device edge is key to speed of development, speed of deployment and scalability.
With applications decoupled from hardware and packaged with their library and configuration dependencies, they can be deployed across a range of hardware platforms, at any level of the stack, from the smart device edge layer to the on-premise data center to the centralized, cloud-based data center. This affords architects and engineers the ability to scale applications quickly from proof-of-concept to global scale, volume production deployment without budget-busting, rip-replace engineering cycles, for either software or hardware.
Containers also unlock the value of open architecture edge compute hardware. Each container can operate independently, eliminating the concern of breaking an existing application by installing a new application due to conflicting dependencies. In a containerized world, use of compute resources is constrained only by the processing and memory capacity, not by the fear and uncertainty of creating application conflicts. This is particularly critical in the world of digital transformation where we fully expect that data analysis and technology evolution will illuminate new use cases over the life cycle of deployed hardware.
Analogous to discovering and unreservedly installing new apps for your smartphone that didn’t even exist last year, leveraging open architectures and virtualization technologies will enable your digital transformation investment to adapt and grow with your business needs, whether they be road mapped or yet undiscovered.
When Malcom McClean set out to eliminate the inefficiency of the shipping industry, he could not have imagined the role that would play in the global trade economy. Software virtualization has roots dating to 1960 mainframes but, with a growing confluence of needs and benefits, container-based virtualization has the potential to transform and accelerate the world of open architecture edge computing, digital transformation and the IIoT.
