Interview at the edge: building the company
Edge computing sounds technical, digital, even mystical, and vague. But people working on the solutions and their projects are very real. Here at Hivecell, the time has come to start telling their stories.
Beginning with Volodymyr Kondratenko, Managing Partner at Ricker Lyman Robotic, and EMEA President at Hivecell.
Hello Volodymyr. Could you start by describing the current Hivecell platform from your perspective in 1–2 sentences?
Hivecell is a complete solution that lets businesses use the same software on the edge that they run in the cloud, thus providing a lot of savings, centralized remote management, and scalability.
But Hivecell started very modestly. Give us a brief history of how the company was founded.
I will tell the story from my point of view. Originally, Jeff Ricker — today CEO of Hivecell — created the idea/concept of Hivecell as a small distributed server in 2015. He even registered a trademark and company all on his funds.
He was searching for the right opportunity to build the product, but at the time continued to work on Wall Street.
We met in 2016 when Jeff was looking for a team of engineers for his employer. I suggested building a company using that team for the initial leap. Jeff, his engineering partner Paul, and I spent a lot of time discussing whether we were ready to create a company together. Eventually, we agreed that we could not lose the chance.
That is how it turned from a one-man company into something bigger. We started in Lviv (Ukraine) with a team of four people sitting in the same workspace for two months until we found a bigger place with two rooms and conference space to move into. Jeff and Paul were joining us online in the evenings — that was a distributed team for the distributed computing solution.
What was the definition of Hivecell as a product when Jeff first contacted you?
The definition of Hivecell at that time sounded like “Stackable servers for distributed software you can place on the engineer’s desktop. They are extremely scalable and easy to start working with.”
Now it is a PaaS.
Why pivot?
We had two pivots.
The first was a shift from desktop scalable servers for engineers to an edge server for business use cases. We were still selling hardware with some software on top of it.
The second was the pivot from a hardware-based distribution to the PaaS model that provides computational power and centralized management at the edge.
I will talk more about the second pivot, as it was much more crucial. The biggest challenge that all companies are facing is not just edge processing (IoT as well) but infrastructure in general, the way hardware and software providers cooperate.
Hardware providers just ship you a box — now you have to deal with it on your own. Software providers give you a list of supported devices, and now you, the client, have to solve the puzzle of putting it all together.
That is what makes us different — hardware is not our core but simply an enabler of the platform format and that frees the client from “puzzle games” and lets them concentrate on business logic implementation, exactly in a way the cloud does — but significantly cheaper with the edge solution.
Cloud services removed the burden of low-level hardware management and its scaling challenges, Hivecell offers the same benefit but with the extra value of significant savings on electricity, bandwidth, and cloud service charges.
How has the ever-changing product surprised you in the previous years of development?
One funny moment came to my mind. We were building a prototype of Hivecell One, and there was a hot discussion on the capacity of embedded UPS. The initial requirement was to provide up to 10 minutes of autonomous work to ensure data safety and correct business logic shutdown. But while the hardware modules were still in development we could not measure maximum energy consumption rates. When we assembled the first prototype and launched it to run GPU cores and full HD video system it was working on the battery for 45 minutes. You should have seen Paul’s face when he heard that.
Can you share a case study of a Hivecell internal use(s)?
I have two that I am an extremely big fan of.
First is our educational project we engaged in together with Ukrainian Catholic University. We deployed a system that consists of six IP cameras, one simple switch, and two Hivecell units, 10 meters away from the cameras.
The task was to provide a testing bed for Computer Science master students, who were running video recognition machine learning models on our set up. These test assignments and pet projects provided the students with real-life hands-on experience in machine learning on the edge. It is just one of the projects we did together with that University, but it really shows the power of Hivecell product — the whole setup took only one day, with 90% of the time spent on mounting cameras and network cables to them, and literally one hour on the server and software setup. It’s that easy to use!
The second case: right after AWS published the AWS GreenGrass framework we wanted to run it internally on Hivecell. In a day we got it all up and running, and we were pushing Lambdas to edge devices in our Lviv(Ukraine) and Iowa(US) offices. Later, the same GreenGrass let us install a Spark cluster on a Hive and push Spark jobs to it. It was a turning point for me in understanding the technical capabilities of our product.
COVID-19 — what makes Hivecell tech actual?
When I am thinking of existing challenges with the virus and maybe potential future ones I first consider how society will adapt to it.
It might be true to suggest that social distancing will become more common, the business will become more and more distributed and decentralized. This gonna lead to obvious spikes of local offices, on-site solution centers. The business will be interested in decreasing traffic between centers and steady communication in the virtual domain.
To be honest, the whole decentralization and distributed structure is the core of Hivecell. So while we worked on solving emerging edge computing trends we have in fact created a model that is fit for a business of the future.
