Building Edge-as-a-Service Business and How to Move to the Edge

Hivecell is an edge as a service company. In recent years and as one of the up and coming services of the tech world, edge computing has started to take the market by storm—now a top priority for most fortune 500 companies and projected to grow exponentially in the coming years. The idea of edge computing and how it works is really quite simple and comes with an array of undeniable benefits for companies’ growth. Hivecell was formed from the realisation that not all software will be able to run in the cloud, leaving tech businesses scrambling to try to solve this soon to be an issue for all. 

Jeffery Ricker, CEO and Co-founder, and his team at Hivecell had anticipated this shift and the need for an alternative to the cloud and so, therefore, built a solution to it. It addresses how to deploy software on-prem, at hundreds or 1000s of locations, whilst keeping it as close to the cloud-like experience for the software developer and the customer as possible.

200+ companies from 25 countries outsourced software development to Relevant

We provide companies with senior tech talent and product development expertise to build world-class software. Let's talk about how we can help you.

Contact us

In this article Jeffrey shares with us: 

• What is ‘the edge’ and how it’s growth is going to be exponential in the coming years. 
• Why companies need to start the transition to edge computing now in order to stay ahead and not waste money and resources on old, archaic solutions to new problems. 
• Which country has, in his opinion, the best engineers and why.
• How telling your team ‘You’re not thinking big enough’ can help produce some of the most incredible results.

Edge Computing: The beginning of a new trend

“It’s the ability to run software outside the data closet, the data centre, or the cloud. It’s running software, in 1000 retail locations, in 100 oil fields, on 100 ships at sea, on the factory floor, or in warehouses. It’s out there, it’s where the data is.”

Jeffrey tells us that there has been a tide shift in recent years and up until only two years ago, more and more data was being generated in the cloud and being pushed out to the edge, out to the cell phones, out to the homes, and out to the offices. He highlights how the majority of the data was generated in the cloud and then sent to the edge. 

“The majority of the data is being generated at the edge outside the cloud and needs to be pushed to the cloud—this is mainly being driven by the Internet of Things and machine learning. Here is this brand new source of data, IoT—a huge amount of data, much more than we ever anticipated. And then there’s machine learning—a brand new way to extract data from that data source.”

One step ahead: How Hivecell predicted the rise of edge computing

“Well, we were first involved in helping companies adopt distributed computing. That is software that’s designed by nature to run on multiple machines. We realised that there was this revolution that had occurred in software, starting with Hadoop with big data, but then being followed by every other major software. Then after that Kubernetes, Kafka, and all the other buzzwords that you’re hearing now. The hardware industry had completely ignored it.” 

“If your software needed more memory, more disk space, and more CPU power, you needed a bigger box, but most applications didn’t really need all that compute power. In fact, that was way overkill.”

“Then came the revolution of distributed computing, in which distributed computing makes several small computers act as if they’re a big one, now enabling you to scale an application by adding more and more servers.”

Jeffrey tells us how this led to an insane process benefitting nobody eventually led Jeffrey and his team to predict the growth and undeniable need for edge computing. 

“We had big boxes that we made look like little boxes, and then we ran applications on top of it to make the little boxes look like one big box. It’s insane. We realised the tide had changed and you’re not going to be able to process all this data in the cloud, but have to push the compute out to the source of the data. That’s where Hivecell comes in.”

Benefits: Who should use edge computing? 

“You can’t run everything in the cloud. If you’re a retail store, you can’t put your point of sale system entirely in the cloud, factories also face an issue of autonomy. This is because you’ve got to run the factory, there are security issues, and the data is just too sensitive to move to the cloud. Another issue is that there are compliance issues, such as personal identifying information, video processing, and bandwidth issues, and they are just too big to move there—leading to cost issues.”

Jeffrey explains how we’re seeing two different reasons for the need for edge computing, but as we are only in the beginning stages of the edge revolution, these reason and the demand for th edge is sure to evolve over time.

“There is so much IoT data, and it can’t economically be moved to the cloud. And so, you have to ask yourself, how do I take advantage of that data? The answer is Edge Computing. You’ve got to bring the cloud to the edge.”

Transition to the Edge: Stop archaic ways of thinking when approaching the Edge

Jeffrey explains how the transition is simpler than many believe, but it does involve a little education and, most importantly, an openness to a fresh, new perspective, as with all good things. As a former Calvary Man in the US Military for almost ten years, Jeffrey naturally likens many situations to those encountered on the battlefield. When it comes to the shift to edge computing, he defines people’s natural desire to stay to the status quo and not think outside the box as the notion of “Generals fight the last war,” which implies when companies are trying to solve a problem, they tend to apply the solutions of the previous challenges without adapting to the current one.

“People try to go back to the way things were back in the 1990s or 2000s, where you buy the hardware, you send a technician out to install it, and you manually install the software there. This way is a nightmare, and it doesn’t scale. It takes a trained technician for hours to install a Kubernetes cluster and if you’re dealing with 1000 stores, this suddenly becomes a big problem.”

“You have to start with a clean sheet of paper, you have to put all the assumptions on the page, and come up with something that’s designed for the problem you’re solving. It’s not by coincidence that we call them software architects or solutions architects because that’s what a real architect that designs buildings does; they start with a blank sheet of paper, and talk about what is this building for and what is its environment. They don’t just lift a building previously designed in one place and drop it in another.”

MVP  development: Figure out what you’re doing and why you’re doing it

“The first thing I would say to co-founders is—it’s not the investor’s problem to figure out what you’re doing and why you’re doing it. You have to work on your message.”

“One of my mentors told me that you’re going to have to pitch 50 to 100 times before you figure out how to actually convey that message. It was a horrifying realisation for me. But it turned out to be true and it’s just the nature of the game. You just have to learn how to communicate your story effectively.”

Exponential growth: To outsource or not to outsource

Co-founders, Volodymyr Kondratenko, built an engineering team, a wholly-owned subsidiary, in Lviv, Ukraine, with the majority of the team being located there. Jeffrey tells us that even in Ukraine, with a wholly-owned subsidiary, the struggle to find the talent is still a huge challenge. Jeffrey highlights how he believes the talent is there to be found but the problem is now centred around how to pick up the talent when the competition is so fierce—tech companies from all over the globe competing against others in the hunt for that world-class talent. Jeffrey tells us that, upon starting Hivecell, it just wouldn’t have been affordable to outsource the talent from the States and so they, therefore, looked outwards. However, over the past five years, there has been substantial inflation in the rates of salaries expected in Ukraine, to the point where other markets, such as Canada are comparable. 

Where to hire: Why Ukraine? 

“There’s something about the education and the environment in the culture of Ukraine that just produces world-class engineers.”

“One of the ways I sum this up is in other cultures, the engineers are very quick to say ‘yes, we can do this’. Leaving you having to push very hard to figure out what are the actual challenges because they brush over them. Usually resulting in the project going off the rails and stagnating, not realising for six months that there are some fundamental flaws in what you’re doing and how you’re doing it.”

Jeffrey tells us how in Ukraine, engineers take a different approach to working on a project in comparison to other engineers he’s worked with—Ukrainian engineers always start with a ‘NO’. He tells us how they start by doubting the realisation of what you are asking of them, telling you how It cannot be done. But, they are willing to work on the proposal, and over time, step by step, they identify all the things that will make the project feasible. Ukrainian engineers essentially work on each point from a position of ‘No, not possible, to a ‘yes’. 

“This only happens when you can have an honest discussion. It’s a much faster path to a solution and allows you to be able to trust your engineering team a lot more. You know what the risks are, what’s feasible, and what’s not. This isn’t something you get if an engineering team just says, yes, and hides that from you,”

How to lock them in: Attracting the talent to come work for Hivecell 

Jeffrey notes that one of the ways they have fought the battle of recruiting the best hires is through investing in the culture of the office in Lviv, investing much in the aesthetics of it. He explains how this helps to create an environment that is better than others, to build a place that you want to work and a team that you want to work with.

He also notes how Hivecell is a very open company that wants its staff to know what’s going on in the company, keeping all in the loop in the strategy and decision-making process. Additionally, Jeffrey marks how all staff receives equity, which is still an uncommon approach for many Ukrainian companies. 

“We do know how a startup can be a risk. So what you’ll find is some engineers and other talents, seeking out that risk, and, yet some avoid it. So there’s some talent, you’re just not going to get because of the risk threshold. Startups have ups and downs. They have crises, they have good times. People just by their nature will thrive in that, some of them will be able to tolerate it, and some of them will not be tolerant—they really fall into three categories.”

Managing offshore teams: What advice can you share? 

“First of all, you’ve got to have somebody there to lead, someone you can trust explicitly and visit as often as you can. Before the war, I was going over at least once a quarter to spend two weeks there. My other partner was there for half the year, spending half his time in Ukraine and then bringing the team over to the States. That’s also a key part of it. There has to be cross-pollination.”

Fallbacks of Edge Computing: Security breaches, vulnerability, physical maintenance, management, time, cost, and connectivity issues

“IT security is paramount in the thoughts of our customers with Edge. And I would say that, again, going back to the blank sheet of paper, the security threat of edge is very different from the security threat of Cloud. It’s a completely different attack plane.” 

“So with Edge, the way we do it, you have 1000 attack points but every one of them is only an inch deep. If you breach that one, you end up with a set of data that’s probably meaningless in value, without any value to whoever’s hacking it. For example, the vibration data off of a dynamo or the transactions of a single store.”

“Of course, our professionals know how to secure a system. It’s not like we just throw all that stuff out the window. Cloud has a very small attack surface. But it’s a mile deep. Once they’re there, they’re in everything. It’s a different way of thinking about security. A single node does not compromise you because they would have to hack every single node in order to do anything. Whereas in the cloud, or in a data centre, once they were breached, they’re on the inside of the firewall. Then it’s an open plane. It just has to be thought about differently. Again, going back to fighting the last war, you cannot just take how you think about security in the cloud or the data centre and slap it on Edge.”

Networking is a very key part of the Edge

“The network that an edge application has to deal with is much different than that of a network that a data centre works with. Nobody would design for a server failing. If the data centre server fails, we have resilience, we have redundancy, and the applications respond. But we don’t have all of a sudden every IP address changing. There’s a lot more to it. It’s a different animal. Let’s say that.”

The transition: When to start utilising the Edge? 

Use case: Here at Relevant we specialise in IoT software development. Recently we had a client that asked for a custom system to manage wind turbines. How should we approach this decision with Edge Computing in mind? 

“First off, you’re not going to be able to just move all the raw data to the cloud and process it there, it’s just going to be impractical. So that means you have to process the raw data at the source, and then send only the business-relevant data to the cloud.” 

“The raw data is meaningless to anyone if you’re monitoring the vibration of a dynamo. At 1000 hertz, there’s no business value to that data. All you want to know is: is that vibrating normally, or abnormally. When you compare that binary number to the vibration data, it’s a huge difference. Plus, you need to be able to know whether you have network connectivity or not, so you may need to shut down the dynamo. But it’s even more important to save the dynamo.” 

“The second thing is that you need to start now. Start small and scale quickly. So, you do not have to put in a $100,000 server at those locations like other companies are telling you. Take data center hardware and put it at the edge. You do not need a 72-core server at the edge, similar to how you have in the data center. You need an 8-core server, and you need three of them.”

“There’s this concept of thinking of edge scale that if you have a 72-core server or a 96-core server, and I have 100 of them in my data center, then that’s a lot of horsepower. But when thinking about the numbers, that’s about 10,000 cores of processing that you have. If you take that same 96-core server and you put one of them at 1000 locations – That’s overkill. That’s so much computing power. But if you put 33, eight-core computers there at the edge at 1000 locations, the engineers that are in the data centre ask, ‘what are you possibly going to do with 24-cores’?” 

“Well, the fact is the loads that you’re dealing with are not concentrated, you’re not dealing with this data of 1000 locations, you’re dealing with the data of one location. And so the workload is not the same as the data center.” 

“If I tried to take datacenter, hardware data center software and put it at the edge, it’s like trying to drive an attack nail with a sledgehammer. When it’s added up, it’s a lot of infrastructure.” 

“So start small. Your edge deployment should be the smallest possible size necessary to do the work. Show that return on investment, and then scale it up. As you add more workloads, you will scale because whatever you’re doing at the edge now is only the beginning. It’s going to grow and grow and grow. It’s going to grow exponentially. So anticipate scale.” 

“Don’t buy that $100,000 server and say, Oh, in five years from now, we’ll have a workload big enough for this. That’s not how any sane company builds return on investment these days.” 

“The customer who assumes they can process all of their IoT in the cloud – it’s like the 1950s sitcom where the dad builds the boat in the garage, and then he can’t get the boat out of the garage.” 

“We’re also seeing companies that say, ‘I can do edge but I’m gonna home grow it, hey, Kubernetes is open-source. I can put this stuff together. They then realise that once they get to the 10th or 20th location, they can’t manage it. They can’t scale it.” 

Military experience and Leadership skills: ‘You aren’t thinking big enough’

“There’s no greater institution for teaching, studying, and fostering leadership than the United States Army,” Jeffrey told us. He speaks of how the military was invaluable, providing him with the understanding of what it takes to be a leader. He highlights the importance of taking ownership for everything that goes wrong and how a good leader should give praise for all that goes right to his team. “It’s uniformity of command, having one person in charge of each problem,” Jeffrey adds.

Jeffrey highlights how these skills have been transferable into the corporate world, effectively helping him succeed, not only in managing his team but also by thinking about how they should go to market and how to deal with strategy.

“I use metaphors all the time which relate to war. I’ll talk about doing a Calvary screen on the market – ‘You’re spreading yourself thin and wide looking for that opportunity. But once you find it, you concentrate everything on that one thing and you push forward.’ You hear people talk about it in startups all the time. But for me, it’s the same as the Calvary screen in a breach. 

As well as serving in the Reserves as a Cavalry officer, Jeffrey also worked in defence research as a civilian, highlighting how it and their leadership there has also had a huge impact on Jeffrey’s success and how he motivates his team.

“I was working on some amazing projects. I was in the supercomputing program for the Pentagon. My mentor, General Minnis, told me, ‘you’re not thinking big enough’. No matter what I brought to him, his response was always you’re not thinking big enough. This mantra of ‘Pushing me to think bigger’ was formed there.

I try to push my engineers, my architects, and my UX designers by saying, ‘you’re not thinking new enough, you’re not thinking novel enough, you’re not pushing it far enough, is there something better out there, is there something else that we could be doing?

I don’t know if there is, I’m just asking the question, I’m just being that thing that makes people think and they always surprise me. They always blow me away with what they come back with. I’m now at that point where I’m not inventing things, I’m creating an environment in which things can be invented, and thrive.”

Challenges: Some challenges in building Hivecell

Jeffrey mentions the importance of balancing the three cultures—you have to balance within a  startup; the engineering—the research and development side, the operations side—beans, and bullets, payroll, keeping the money, and the sales and marketing side. Jeffrey admits that in the beginning, the approach to building Hivecell was a little lopsided as he and both his co-founders are all engineers by profession and nature. 

“The biggest challenge we’ve faced is being weak on the sales and going to market side. There are three different cultures and they’re like three legs of a stool. If one of them’s too strong or too weak, you’re going to fall over. The key here is to reach out and find the partners that have the skill sets that you need to shore up your sight of the weakness and trust them to do what it is they do.”

“The single the largest challenge of any startup is finding enough competent leaders – If you can find the competent leaders, you will create a competent team, because competent people will not tolerate working for incompetent leaders, they’ll go someplace else. Everyone leaves a job because of their boss 100% of the time. If I could find 12 competent world class executives, I could conquer the world. But Jesus Christ himself couldn’t get 12 competent. So it’s to find that team that you can trust, that’s competent, and really take care of themselves.

Jeffrey believes that the reason why it’s so hard to get those 12 leaders is because anyone that is competent is capable of starting their own company. 

“It’s so getting people that are willing to join forces, each one of them could be running their own company, but getting them all to join forces in a single company. That’s the challenge.” 

Your Next Read