Edge Executive Interview – Carlo Daffara, NodeWeaver
In the lead up to Edge Computing World, we’re taking some time to speak to key Executives from the leading companies supporting the show. Today we’re talking to Carlo Daffara, Founder & CEO at NodeWeaver.
Tell us a bit about yourself – what led you to get involved in the edge computing market?
I have been involved as a researcher in many EU-funded projects that were focused on helping small companies and public administrations in making use of open source to create resilient and useable IT infrastructures; one of the outcomes of that was a platform designed to run applications in a self-managing way on limited hardware, in a sense a precursor of edge computing. When we got the opportunity to turn this into a company, I decided to go and work full time on this; we had a major industrial group as one of the first customers, trying to provide high availability to plant management in a very harsh environment. Being an entrepreneur is quite different from being a researcher in European projects; we have been working as a company on this for nearly 10 years now, and it has been an incredible experience.
What is the company uniquely bringing to the edge market?
NodeWeaver is a bit of an anomaly in the market- most solutions are designed to be managed from a central management system, and most are unable to work properly without connectivity. We have been working with customers in remote areas where connectivity is not always available, so we have been working from the start with the idea that the edge should be able to work and operate autonomously. This autonomic capability is probably the key differentiator- all aspects of the hardware, the fault management and the tuning and optimization is entirely AI-powered. We are also extremely lean – we provide full software-defined storage, networking, orchestration in systems as small as 4GB of memory, one order of magnitude smaller than most of the platforms on the market and also the only platform which provides these capabilities on both x86 and ARM devices – all while being able to handle tens of thousands of “nano-clouds” anywhere in the world.
Tell us more about the company?
We started as a spinoff, and we had the opportunity to work in private beta for several customers that adopted our solution and helped us make it better. Being funded to research the applications of edge computing in many different verticals allowed us to invest in making the platform perfectly adapted to the needs of industrial computing, telcos, smart cities and many others, without having to go to market immediately. This allowed us to go through many iterations, especially for what concerns usability at scale (reconfiguring software while on the top of a telephone pole is not easy!). and we now have nearly one hundred deployments in 15 countries. We had been lucky to find a way to grow organically, and our team is made of extremely talented (and experienced) people.
How do you see the edge market in Europe?
The edge market is on the verge of exploding- thanks to the clear value of the use of local high-efficiency processing, especially in manufacturing (a European strong point). The use of deep learning at the edge, capabilities like anomaly detection, secure management- all point to a major adoption. The increase in the number of sensors and data sources make impractical to send all data to the cloud for processing- we have customers that would require up to 40Gbit/s from each remote manufacturing plant, and that is economically unfeasible. Processing at the edge, sending to the cloud just a small percentage of the relevant data is a simpler and cheaper solution, and companies are adopting that at an accelerated pace.
How do you see the edge market developing over the next few years?
We are now mostly in a phase where vendors are exploring different tools and methodologies, coming from different backgrounds. A startup may approach the market by providing a very specific vertical solution, while larger vendors may try to create horizontal platforms and then leave verticalizations for others (or the customers themselves). Some standards are emerging, but this is really like the initial cloud period- exceptional growth, lots of confusion and lots of different solutions that will consolidate in a short period of time into far fewer market leaders. There is still much work to do to make the solutions simpler and especially more robust for large scale deployments; many of the tools that we have seen are fine for a small lab deployment or a proof-of-concept, but scale badly when you have to deploy them in hundreds or thousands of places. This also explains why some of the initial experimentations did not get too far; I believe that we have created an excellent solution to cover both the task of managing the deployments and for the applications on top.
Thanks Carlo – Looking forward to hearing more from NodeWeaver at the event !