- Jason Shepherd
Tackling the Biggest, Smallest Things - Why I Joined Nubix as CEO
Throughout my career, I have tended to find myself on the very front end of emerging trends. Accordingly, one of my many mottos is “if it’s fuzzy, I’m on it!”. My journey in the IoT space began in late 2014 when a small group of us in the Dell CTO group formulated a strategy that was eventually funded as the company’s IoT business. I then led a team to build out our partner ecosystem and enable solutions on top of the company’s edge hardware.
I spent much of 2015 doing what I dubbed the “Magical Mystery Tour” - visiting myriad startups offering IoT platforms to find the right lead software partner for Dell. Many of these platforms had been adapted from the M2M days, and the typical edge “gateway” component was architected as a monolithic block. On one particular trip, I had an epiphany that may be nobody was doing it right quite yet. A call from the road to my Dell CTO colleagues started an incubation effort that led to the 2017 formation of EdgeX Foundry in the Linux Foundation.
Clearly playing off the name “Cloud Foundry”, the idea behind EdgeX Foundry was to extend cloud-native principles (e.g. platform independent, microservice-based architecture, CI/CD) to the edge, in addition to facilitating interoperability between heterogeneous components through a unifying service and API layer. Thanks to the help of a vibrant developer community, the EdgeX framework is now being leveraged worldwide to facilitate the development of modular, interoperable IoT solutions.
In 2018, I was recruited to join the executive team for the newly formed Dell Technologies’ Edge and IoT business that coordinated efforts across the portfolio, including Dell, Dell EMC, VMware, Pivotal and RSA. As the group’s CTO, I led a diverse team of industry SMEs that continued our edge ecosystem development spanning verticals like manufacturing, energy and retail, curated emerging technologies such as data confidence fabrics (which led to the 2019 formation of Project Alvarium in LF Edge), and built a strategy for cloud-native orchestration at the distributed edge. Much like EdgeX does in the application plane, the idea was to extend cloud infrastructure principles into the field.
In 2019, I decided it was time for a change and left Dell for ZEDEDA. During my three-year tenure there as VP of Ecosystem, I continued with my passion for growing the network effect at the edge, in addition to serving as governing board chair for the Linux Foundation’s LF Edge umbrella organization. I was an ambassador for the LF Edge projects and a significant contributor to the community’s 2020 and 2022 taxonomy white papers that break the edge continuum down based on inherent technical tradeoffs
The 2022 LF Edge white paper specifies four overall edge management and orchestration (MANO) paradigms: Data Center, Distributed, Client and Device. While the ultimate goal is to leverage cloud-native principles as much as possible across the edge continuum, inherent technical tradeoffs necessitate different underlying MANO tool sets for each paradigm. I mean, it’s not like you can plop a full-fledged data center stack down on a Raspberry Pi. Footprint aside, management and security solutions for edge data centers are quite mature, as are those for client devices like PCs and smartphones.
However, leveraging cloud-native principles for the other two edge MANO paradigms is still an emerging concept. ZEDEDA, among others, is focused on distributed edge nodes that are outside of the data center but still capable of running Linux. This includes gateways and small server clusters running OCI-compliant containers, virtual machines and Kubernetes, and are typically deployed in locations such as manufacturing floors, oil rigs, retail stores, vehicles, and more.
Meanwhile, it has pretty much been business as usual for edge devices that are powered by highly-constrained microcontrollers (MCUs). Development for MCUs requires specialized skills and is time consuming, and the resulting software is monolithic in nature… much like the IoT gateway stacks that I saw during the 2015 Magical Mystery Tour.
Given that Arm estimated that two-thirds of the 160+ billion edge devices in 2019 were based on microcontrollers, they make up the biggest, smallest edge footprint out there. As such, there is a massive opportunity to streamline the development and deployment of solutions that leverage these devices.
In late 2022, I took a short career break to recharge and then did a little consulting work on the side as I considered my next adventure. I was certainly intrigued when the Nubix team approached me to join the company as CEO. The solution is a natural extension of the work I’ve done in the edge computing space over the past eight years with the help of a bunch of great people.
Nubix’s mission is to extend cloud-native principles from the upstream “edges” to the billions of microcontrollers in the field. The bare metal Nubix runtime supports tiny containers on edge devices which abstracts applications from the complexity of embedded hardware. This enables faster time to market, in addition to a broader development pool since you don’t have to be an embedded specialist to build apps on top of the Nubix foundation. The Nubix Orchestration Hub enables developers and admins to remotely deploy and manage their edge devices at scale.
The solution also provides benefits such as fractional, lower risk app updates, increased security, and more. Plus, there’s a significant benefit for IP protection because you can obfuscate your proprietary code (e.g. a Tiny ML model) as a containerized app, instead of having to expose your raw source code to a third-party developer so they can compile it into a traditional monolithic image. Nubix’s special sauce is providing all of these benefits while minimizing the overhead for edge devices in terms of footprint and power consumption.
It’s an unprecedented time in the technology world with the pace of innovation ramping at an exponential rate. Despite all the recent focus on edge, the cloud certainly isn’t going away. It’s just that applications need to be more distributed in nature due to the sheer volume of devices and data on networks.
While the typical justifications for edge computing are technical in nature, such as low latency, conserving bandwidth and ensuring autonomy, security, and privacy, another key driver is customers starting to pull back from their investments in public cloud. After all, the “easy button” of the cloud sounds great until you get the bill. Edge is as much about users taking back control of their data as it is the technical reasons.
As such, we’re seeing a rapidly increasing focus on multi-cloud strategies which includes reducing dependence on infrastructure and tooling provided by any given public cloud. Key to this is decoupling software from hardware, leveraging open source, and breaking dependencies into interchangeable modules. Nubix’s approach enables this strategic imperative for constrained edge devices by abstracting cloud dependencies as containerized applications, instead of being hard-coded into a monolithic binary.
I’m excited to carry the torch as Nubix’s CEO as we simplify development, scalability and security for MCU-based edge and IoT solutions. Kudos to Rachel Taylor for leading the company to this point as CEO, and I wish her the best in her next adventure. Over the past six months, the team has brought on key talent including Stephen Beard as CTO and Brian Dunlay as lead engineer. We’ve commercialized the platform, continue to refine it for greater efficiency and lower footprint, and are starting to build out our ecosystem of hardware, applications and solution providers.
Don’t hesitate to reach out if you’d like to learn more about how we can help you accelerate and scale your edge or IoT project, or how as a developer you can get involved in shaping this groundbreaking technology. The future is bright for Nubix as we bring the agility of the cloud to the biggest, smallest edge footprint out there!