How to ride the ‘edge-to-cloud continuum’

Jos Schuurmans 26-02-2021 Energy Industry 4.0 Smart society
Photo by Jörg Angeli on Unsplash.

Photo by Jörg Angeli on Unsplash.

We talk quite a bit about the ‘edge-to-cloud continuum’. Fancy term, eh? It’s actually a thing and part of the evolution that RAIN enables.

Our timing in taking RAIN to market is no coincidence. Its technology plays an essential, enabling role in the current data revolution. To understand why our platform is so groundbreaking, it is helpful to see it within the context of the decades-long evolution of Information Technology.

In short, RAIN makes it really easy for anyone working in an organization that takes its data seriously, to build data applications by dragging and dropping a data feed onto a canvas, connecting the feed to some routine that manipulates the data in some way, and then adding an action of some kind to the outcome of that manipulation.

Connecting these elements - input, processing, output, and action - is what we mean by creating a data application on RAIN. Each of these functional elements, or ‘nodes’ as we call them, can be re-used and re-combined in order to create new applications. How cool is that?

Once you get the hang of it, you don’t want to stop.

What makes it even more beautiful is that these data applications can be deployed on any connected hardware throughout an organization’s infrastructure. RAIN is a horizontal software platform that can be installed on any operating system, or across any mix of operating systems: on devices at the edges of the network, on cloud services, and on any computing resources in-between.

Through RAIN’s web UI, one can connect any data feeds coming from any connected machine. The physical location of where the processing node runs can be anywhere. The output can be shown anywhere. The resulting action can take place anywhere. It all happens on the edge-to-cloud continuum.

An evolutionary tale

Sounds and looks simple enough, right? That’s what makes RAIN ‘no-code’. But the development of all the underlying technologies, the tech stack, the IT architecture, did not happen overnight.

Our Head of Product Pekka Immonen made a real effort to explain to me how we got here. The IT architecture illustrations in this blog post are from a slide deck that he made. I will try to unpack things the way I’ve understood them from Pekka.

No-code enterprise data applications on the edge-to-cloud continuum are a far cry from how IT systems started out.

In traditional IT software applications, we typically have a database to/from which data is written and read. Data is manipulated in a process layer that follows a business logic, and the outcome is shown to humans in a presentation layer. Sometimes multiple manipulation and presentation layers use the same database. If the business logic software is upgraded, it has to be installed on every instance.

Compare this to the edge-to-cloud continuum, where a data processing node is maintained and developed in a single instance. When upgraded, it is immediately available to every node and application (ie. combination of nodes) in the system. The RAIN application is so modular and flexible that only the latest version of a node is available and called upon.

In fact, the modularity and flexible availability is much of what Cloud-Native Computing is about.

Cloud Computing is sometimes somewhat primitively seen as moving computing resources to the cloud. Migrating server software from centralized computers in-house to virtual, managed servers in a cloud hosting environment doesn’t quite make the whole setup cloud native.

Cloud-Native Computing deploys so-called microservices that are maintained and developed independently.

As the Industry 4.0 paradigm with its IoT and IIoT variants has been picking up steam since the early 2010s, the direction of data from sensors in industrial equipment has very much been one-way: from production machines to a computer inside the factory initially, then to a cloud server, and ultimately to a cloud-native environment.

By bringing computing power closer to sensors and other devices at the edge of the network, significant cost reductions can be achieved. Instead of constantly sending all the data that an air quality sensor collects to the cloud, with Edge Computing it is possible to process that data close to the source, even run an artificial intelligence algorithm to detect patterns and make predictions, and only send the result of that computation to the cloud.

With the number of data sources growing exponentially, Edge AI is the only feasible way forward. The internet was designed for downlink data traffic; not uplink. Transferring all un-reduced, raw data to the cloud for processing is physically and economically unsustainable.

As long as the operating systems of industrial embedded software are proprietary and diverse, true interoperability between the edge and the cloud is hard to achieve. Once we do achieve such interoperability, we are talking about Cloud-Native Edge Computing.

It requires a software layer that can communicate with all the various operating systems to harmonize them and make them interoperable. This layer, which is also referred to as Orchestration software, makes the nature of the different underlying operating systems invisible to application developers and users.

Kubernetes is an open-source software solution, originally developed by Google, that is fast becoming the de facto standard for edge-to-cloud orchestration. It actually creates the edge-to-cloud continuum on top of which RAIN can be deployed.

Our CEO Henri Kivioja made a bet this week that “95 percent of all companies will use Kubernetes within the coming 4 to 7 years”. Now that’s a rather bold claim, isn’t it? I didn’t quite dare to ask Henri what he’s betting, but I sense that it’s the farm.

The challenge with Kubernetes is that if you want to deploy it on your enterprise infra, and then develop, run and orchestrate distributed software applications on it, you need black belt ninja IT skills like in Pekka’s team.

Now, while that may sound like a business opportunity for us, we have come to the conclusion that it would be nicer and more scalable to develop an all-purpose, easy-to-use,  edge-to-cloud enterprise data application platform that can run on top of Kubernetes (or other orchestration middleware) and hides all the underlying complexity.

Hurray! Let it RAIN!

Did I mention that it’s easy to use? We offer RAIN as Software as a Service (SaaS), so all you need is a web browser.

Say, you want to deploy RAIN to start creating your own enterprise data applications. This is how it goes down.

When you come aboard through our SaaS, we provide you with your RAIN cloud tenant, where you can then start adding your edge devices. Your edge machines need to run an operating system and orchestration middleware like Kubernetes. (If you talk with us, we can help you with that.)

Find the low-hanging fruit

You may have to download a package from our website to install RAIN on your edge hardware.

Once your edge devices run RAIN, they, or rather the data they produce, become available to you as nodes in your RAIN SaaS user interface. Et voilá! Connect some data flows with functional nodes and create your first data application!

OK. So far, so good. You may wonder about low-hanging fruit. So do we! :-)

We discuss with corporate executives every day about the data they use (and don’t use!) to optimize their business. Utilization and optimization is the name of the game, and it’s our bread and butter.

It’s time to start experimenting with data. Together, surely we can identify where to start. Just book a call with Henri.