Nubix Adds Developer Edition for Container Edge Platform
Nubix today launched a free Developer Edition of a container platform optimized for building and deploying applications on edge computing platforms based on microcontrollers.
Tom Yates, vice president of products for Nubix, says the platform employs a “tiny” container architecture developed by Nubix and measured in kilobytes, which are small enough to process data on the endpoint versus rival approaches that require organizations to process data on a gateway or in the cloud. That latter approach inevitably introduces latency because they require developers to move data off the endpoint to process it, he says.
The Developer Edition of the company’s namesake platform comes with support for Raspberry Pi and BeagleBone platforms and includes access to a library of sensors, analytics and services optimized for the Nubix-developed container engine. Rather than asking developers to develop applications in C or C++, developers can access the Nubix platform using open source languages such as Python and several pre-packaged drag-and-drop functions to enable developers to create, for example, an internet of things (IoT) application in minutes, Yates says. That’s critical because a lot of developers building these types of applications need to be able to work in higher levels of abstractions because they don’t have expertise in lower-level programming tools, adds Yates.
Other capabilities embedded in the platform include Tiny Scripts, Tiny Analytics, Tiny Data Services and an instance of an Apache Spark runtime for the platform dubbed Stuart.
Yates says Nubix containers are roughly 100 times smaller than Docker containers, which means it’s now possible to build applications that run in near real-time on the network edge, rather than having to rely on hybrid approaches to edge computing.
For the most part, he adds, the tiny containers developed by Nubix are being deployed on bare-metal platforms largely because endpoints don’t have enough memory to run traditional virtual machines.
It’s still early days in edge computing. However, with 5G networking services on the horizon, it’s only a matter of time before more application code moves out to the network edge. The issue is that many of the applications envisioned for the edge need to be able to respond to events in real-time, rather than waiting on a gateway or a cloud service to process data at rates of latency that will at best be highly inconsistent based on the amount of network bandwidth available at any given moment. The challenge is going to be finding a means to process that code that is lightweight enough to support the processing of analytics to enable things such as autonomous vehicles.
As is the case with most things, when it comes to software innovation, there’s a natural chicken-and-egg tension. On the one hand, developers are clearly excited about the potential of edge computing. However, unless they have access to tools that make it practical to build prototypes of edge computing applications, it’s not likely that excitement is going to turn into anything sustainable anytime soon.
