Accelerating the Intelligent Edge
Updated: Oct 7, 2019
By all accounts, IoT and digital transformation and all the promises of improved OT that came along with them are running behind. IoT was predicted to be $11T a year in revenue by 2025, which would account for 11% of the worldwide economy. The reality of 2019 is that IoT, revenue was estimated to be about $357B in 2019* (IDC) -- that’s just a bit off-track to get to that $11T. There are many factors contributing to the setback including the delay of 5G, the complexity of the IoT ecosystem, the lack of edge standards, and more. Today I’m going to delve into a few of the challenges that have spurred our thinking here at Nubix..
Close Is Not Close Enough
Before I go any further, I want to talk about what we mean at Nubix when we say edge. It’s still early in edge and it’s a term that is not well defined yet, in fact if you ask 10 “edge” companies to define the edge, you’ll get 12 different answers. Most will define the edge in relative terms, they will say their technology is “edge” because it’s farther from the cloud than other options out there. Many companies are talking about putting analytics “closer” to the device, but that’s not close enough to address every use case. For edge to reach the tipping point, we have to be able to address every use case. Being able to address 30%, 40%, even 60% of use cases is not enough to drive enough adoption to reach critical mass. At Nubix, we define edge in an absolute sense, wherever the data is being generated by the sensor/device is the edge, no transport required. Now, let’s talk about a few core requirements to address use cases at the true edge.
The Technology Gap
There has to be an easy way for sensors, devices, data, and systems to easily talk to each other, make sense of what is happening and take action. and all of that has to be able to happen at the edge in real-time. When I say edge here, again, I mean the real edge where data is generated, not the cloud or some near edge gateway. There’s a ton of un-utilized processing power at the edge that can be used to make decisions faster and more reliably. The development infrastructure has to improve by leaps and bounds as well. Cloud development is replete with productivity-enhancing methodologies built around DevOps, containers, and easy-to-use languages and tools. But these are nowhere to be found in edge development circles, which still use the same C/C++ toolkits used 10 and even 20 years ago.
But, what if you had tiny containers -- really, really tiny containers and really, really tiny services -- how would that change your development and deployment processes? Nubix containers are 1/100th the size of Linux containers, which means we can embed our containers on low-power, low-footprint devices. Instead of taking containers and services built for the cloud and trying to shrink them down and make them small enough to fit in near-edge devices, we built edge-native tiny containers and tiny services that fit in as little as 32KB of RAM. Yes, KB not MB. Our containers also provide a hardware abstraction layer that allows for dynamic updates without the potential of bricking the device.
It’s Not Just a Technology Gap, It’s a Skill Gap Too
For the edge to have any chance to reach its potential, it has to get much easier for all developers to build and deploy applications to edge devices. To engage the level and quantity of talent required to build applications and analyze the deluge of data that’s going to be generated by the 50B connected devices deployed by next year, there has to be a platform that allows developers to build applications in the cloud that can be deployed to physical hardware and managed remotely as required.
Our tiny services allow enterprise developers that aren’t familiar with embedded toolkits to write applications for embedded devices. Our tiny services make it easy for application developers to do everything from collecting data from sensors to analyzing data and inferencing with machine learning algorithms to curating data in integrating with cloud-based data services. Complex applications that read data from a sensor and analyze it can be built in just minutes.
A Bridge Over Troubled Water
My team and I are very happy to announce the launch of our Nubix Developer Edition, bringing tiny containers and tiny services to the edge. We have simplified the development, deployment and management of applications at the edge by bringing together hardware, sensors, libraries, protocols, data science and actions -- all through a cloud based portal with drag-and-drop functionality cloud developers expect. Developers build their applications in an easy-to-use interface using open source libraries, then they can deploy it to the container runtime that’s already installed on hardware in the field.
This technology enables a rapid build-test-deploy cycle that is core to addressing the tremendously dynamic requirements at the edge. It also loosens development from the underlying hardware, enabling a Develop Once - Deploy Everywhere capability that companies like ARM realize is core to creating business value at the edge.