Distributed Computing as a Catalyst for Change in the Industrial Sector

We talked about distributed computing with Alex Cantos, VP of Product at Barbara IoT about the great challenges of the Connected Industry and Distributed Intelligence. According to him, decision making in Industry is centralized due to the technological challenge of distributing it, but this is going to change; we have the technology that allows us to bring artificial intelligence to plants and create cognitive machines.


Trained as an Electronics Engineer and Master in Satellite Technology with an Executive MBA from IE Business School, Alex is our VP of Product with extensive experience in the industrial sector as a systems developer in railway and communications environments. He has been project manager in the aeronautics, defense and security industry and director of the robotics area in a critical software consultancy. He is also an adjunct professor at IE Business School. He is a devoted humanist who believes in technology as a tool that should be at the service of the user and never the other way around.

Work ethic:
KISS (Keep It Simple Stupid)

Latest book:
Zen Mind, Beginner's Mind

Music for work:
X&Y by Coldplay

Favourite film:
Blade Runner (the original)

How you define yourself:

Utopia or Dystopia:


Q1. What is your role and mission at Barbara as Product Manager?

I am the strategic product manager, i.e. the one in charge of conceiving and realising products that create value for our customers in a sustainable way. In other words, always being a couple of steps ahead of what our customers need and having those tools ready for when they need them. We also have to evangelise a bit and help them discover new ways of doing things, although sometimes this is not easy.

Q2. How has the Industrial Sector changed and what developments have taken place in your opinion?

The world has changed a lot and yet the industry has changed very little. The world is moving towards decentralised decision making, we see this in many technological innovations and concepts such as Blockchain, Crowdsourcing, Crowdfinancing, Distributed Computing, etc... meanwhile, the industry remains the same.

One of the problems of the industry is that it is based on technologies that are 40 years old and, above all, on paradigms that are more than 100 years old.

When I say technologies, I am referring to SCADAs and PLCs that made sense in their day because they were robust and because they came from a time when everything was done by hardware. And when I mention the current paradigm of industry, I am referring to the approach of Industry 2.0, 3.0 and if you like, Industry 4.0, as they all draw from the same source. They drink from the idea of improving existing production processes, improving the steps of those processes, making them faster, more automated, more efficient, but what we really need to do is to throw the process away and create new processes. New concepts.

And by creating new processes I mean changing the paradigm of production in industry, the paradigm of electricity distribution, the paradigm of water management. Water management companies, for example, are not water managers but data managers. They operate a data platform that allows them to match two forces: supply and demand.

In the energy sector, we talk about smart grids or smart networks. Smart Grids and a marketplace where all actors in the value chain exchange data and can act on the grid. When we talk about the industrial sector, we talk about intelligent and autonomous machines that decide for themselves how they should be managed. In this new paradigm, the operator stops doing manual things and becomes the manager of these intelligent machines.

In other words, it is about rethinking the process from the outside and not from the process we are in, because there is a huge gap between the industry and the technological and social reality . While consumers are used to receiving their parcel almost seconds after completing an online order and knowing where it is at all times, workers in the water industry, for example, see it as normal not to have readings from thousands of meters in their cities for days on end.

Our vision is that this data should be on the mobile device of any employee in the company in real time. What's more, it should make it easier to make decisions, or at least collaborate on those decisions, from that same mobile device. Until we do that, we will not have brought the industry into today's technological world.

Industry today is focused on production, on performance, and what we propose is an industry that is focused on the operators on the shop floor, the people who are responsible for the operation of the equipment, of the machines, who are the forgotten ones in the industry.

We have to put the operator back in control of the management of the machines, give them back the values they lost with Taylorism and put them at the centre of decisions, which is what is expected of a modern industry, only in this way will we achieve a more efficient and therefore more sustainable industry. In this sense, our vision is to humanise the industry.

Human beings today are still doing very basic things in industry and where they really add value is in tasks where abstraction, creativity and innovation are required; in being able to hybridise concepts that are far away from each other, they don't add value by tightening screws. What we see is that people are still carrying out tasks that are not at a high level and this needs to change, and can be changed.

Q3. How developed is the technology to run Distributed Computing?

The technology is already there. We said earlier that the future is decentralised. At a technical level this translates into an infrastructure that manages Edge Nodes, an Edge platform that enables and governs Distributed Computing. That's what Barbara's technology and the products that make up our platform for the Industrial Edge do.

We have created the platform that allows thousands of intelligent Edge Nodes to be managed remotely. These Edge Nodes communicate with each other and make decisions in a consensual and improved way through Artificial Intelligence techniques.

Q4. How do we intend to change the centralized model of work in the industry? What is Barbara IoT's proposal?

Decision-making processes in the industrial sector are slow, they maintain a hierarchical decision-making model with data that are siloed and separated. We are talking about decision processes that are too centralised, all of these are problems in the industry at the performance level and often, there is the judgement of the expert, which today in many cases should not be there, because decisions have to be based on data. The industry must move from intuitive, expert-based decisions to informed, data-driven decisions.

Centralisation in technical or technological terms means that all information flows to a single point and decisions are taken from there.

What we propose is to take that decision to the Edge, to process the data close to the equipment, the machines, the nodes, in a distributed and collaborative way between these "Edge Nodes".

The fact that the industry is not more technologically advanced means that we still believe that there is a manager who knows everything, when it is the people on the shop floor who really know what is going on.

The problem is that there are 200 operators and they don't have a "way" to make decisions with all that knowledge distributed across the plant. Imagine if those 200 operators were Edge Nodes that could communicate with each other and orchestrate that distributed decision making. And that is what we propose in Barbara.

What happens now is that nobody can refute the manager's decision because there are no mechanisms to answer him; and we continue to make decisions in a hierarchical and centralised way. What we propose is an intelligent interface between the operator and the machine. It is the machine that will contain most of the knowledge, elevated by the operator's contribution, which will be contrasted with other nodes to make decisions in a collaborative way.

People talk about Smart Manufacturing now, the reality is that they call anything smart. For a SCADA, what is smart, is it uploading data to the cloud and from the cloud sending one or two commands to the SCADA? To be really smart the SCADA should think for itself, talk to other SCADAs and make consensus decisions.

Our vision is that these nodes collaborate like a swarm, like a school of fish. There is a collective intelligence, which is super efficient and which we enable by distributing and managing distributed computing through Edge Nodes.

Q5. When choosing a technology for the Edge, what would you say are the most important parameters for a company to consider?


The first is to be flexible. For me it is that a product can be very easily adapted to different situations, to different end solutions, to different industries and the ease with which the same product can have different lives. This concept is very much linked to the next variable that is important when choosing a technology, which is interoperability.


Interoperability means that we are able to relate to many other systems. In other words, we understand and speak many languages or, in other words, we offer powerful APIs to users and, at the same time, we are able to access a wide variety of APIs.

There are many ways to cultivate interoperability, but one that I would like to highlight is our connector library, which allows our devices to communicate with most of the industry's most widely used protocols.


In our technology stack we always use standard, open technologies, we don't reinvent the wheel. We are very standard and very interoperable and this is very powerful because it means that our solution can be integrated with any ecosystem and, furthermore, if new technologies appear, it is very easy to adopt them because all these pieces are already designed to fit together.

Our technological proposal is that of a verticalised, highly flexible solution, where cybersecurity is a value.

Q6. What would you say is the main problem for companies when starting an Artificial Intelligence project in Industry?

Many companies know that they have to start applying data analytics, algorithms, etc., in their decision making. They realise that their competitors are doing it and rush to hire a couple of "Data Scientist" who start creating data models in their lab. Algorithms based on the company's business knowledge, as well as historical or data they start capturing. If they manage to create models that help them improve their results, there comes a time when they want to deploy that innovation in the real world, and this is where they run into big problems.

That innovation that has been developed in a laboratory has to be scaled up and deployed across the entire electricity grid, for example, or across the entire water infrastructure, and it is here in mass deployments that the vast majority of companies fail.

And that's because it's far from simple. The Edge is complex, we are talking about a multitude of machines with a multitude of communication protocols, we are talking about making it cybersecure. The deployment in the field, the scalability, is underestimated, and we have the technology that allows these models, algorithms, to get to where they need to go, that is, we are able to distribute intelligence and also do it in a cybersecure way.

A good Industrial Edge platform makes the difference in going from 0 to 100, in turning an innovation into something that delivers results. That's what we've been doing for the past 4 years. We have developed the platform for the Edge that helps companies deploy and govern distributed intelligence in a scalable way, i.e. deploy and run algorithms with very few clicks in many geographically dispersed locations and manage them centrally.

Q7. If the big challenge is to distribute intelligence massively and securely, how do we do it in Barbara?

Our product portfolio is designed for profiles that want to both develop, operate and maintain a network of Edge Nodes. We have created a suite of solutions that meet the needs of different customer profiles.

For the client developer we have Barbrara Edge Orchestrator which allows us to deploy any application or algorithm on virtually any edge node we want.

The user develops their application in their tools and then comes into our environment to deploy and manage those models or those algorithms or those applications on the Edge.

Also, that developer needs their devices to talk to industrial equipment and we have the industrial connectors for that , but they also generally need their applications to talk to some systems in the cloud and we have the cloud connectors for that. Whatever interoperability needs you have, we offer them through different types of connectors.

Then there is the "Operator". Once the system is up and running, the Operator wants to know quickly, accurately and conveniently how the system is doing; which can be materialised in a dashboard where he sees a summary of the status of the entire deployment and, from that dashboard, accesses more detailed information.

There is also an alert system that, when a special event occurs in the system, notifies someone who has that Operator role. We have all of this within a product we call the Barbara Control Room.

The developer can also access the Barbara Biz Builder which is aimed at the type of customer who needs to set up a comprehensive business solution, starting from scratch. The Barbara Biz Builder allows you to set up your own business solution quickly and securely from scratch.

We have users who want to make the whole application from scratch, others who need us to help them make their "stuff" smarter and do it with us, and others who already have the application done and just want to deploy it massively.

For many companies Barbara is a beacon that helps them to see the possibilities of digitalisation. Industry wants to digitise their business, but they don't know how. What industry wants is to "operate", to be productive, and we are that someone who helps them in those first steps of digitalisation, without going crazy with the technology, scaling up, little bylittle.

There is a leap of faith point as well. We ask industry for a leap of faith because software is the devil for them, it's super unpredictable, but the range of possibilities it gives you is enormous. You go from doing three or four things to being able to do millions of things. If you put the risk vs. reward on a scale, the reward far outweighs the risk they take.

We know that in the industrial environment, cybersecurity is one of the barriers to adoption of IT technologies and that is why we have conceived Barbara's platform with cybersecurity by design, right from its conception. And we advocate true IT/OT convergence. We understand the risks of connecting industrial devices that have always been isolated to the cloud, but we have applied security standards and best practices,

If industrial companies want to go digital, they will have to join forces with companies like ours, IT experts, who advocate an understanding with OT. For us, convergence happens on the shop floor, machines are not replaced, legacy machines are maintained. We speak all their protocols and virtualise them.

Q8: Being a technology startup forces us to continuously innovate to create new and better solutions. How does Barbara handle innovation at the product level?

We have a clear 10-year vision, but then we have a short-term vision. We make the product roadmap for 1 or 2 years and we articulate it together with the client; hand in hand with the client, always trying to be one step ahead of what the client expects.

We always start by understanding the problem very well, not so much in providing solutions as in understanding the problem. We do not "marry" ourselves to any solution, but rather we provide solutions to these problems in the most creative way we can. This is how we build our products. We are agile, we have the ability to try things very quickly, discard what doesn't work and persevere with what works.

Our organisation is a reflection of this mentality. Our production branch is divided into three parts: product, engineering and service. The product team is responsible for specifying new functionalities and improvements to existing products, as well as conceiving new product ideas.

The engineering team, on the other hand, is in charge of realising all these functionalities, creating the product, maintaining it and improving it.

Finally, the service team is the one that tailors those products to a customer's specific needs. They are, you could say, the ones who cover the last mile, the ones who make sure the customer gets the solution they need. But the process doesn't end there, all the customer feedback captured by both the service team and by sales or marketing flows back to the product team to further tailor the specification to the needs of the market.


If you were interested in this interview and the topic of Intelligence at the Edge, we encourage you to read the following articles:

Low-code: how to accelerate distributed intelligence at the Edge.

How to deploy Machine Learning at the Edge and what is MLOps

AIoT: the perfect fusion of the Internet of Things and Artificial Intelligence