Making Sense of Distributed Ledger Technologies in Smart Manufacturing Applications

I was recently asked by a colleague to participate in an open-ended interview about the possible uses of DLT (Distributed Ledger Technologies) within the framework of Industry4.0 solutions, and this gave me the opportunity to collect my experiences.

In fact, over the course of almost 4 years, starting in 2017, I've been investigating the impact that several technologies had in the Industry4.0 landscape, specifically finding ways to harness them and create new business model opportunities. DLT was one of them, together with the usual suspects, cloud, edge computing, AI and ML. Even though we had plenty of successes with the latter ones, DLT offered me though challenges and I kept bashing my head against it.

In this article I will present some of my thoughts and findings that accompanied me throughout this journey.

Setting The Stage

First off, I'd like to define a few things. I always preferred to use the denomination DLT in lieu of "blockchain" because I find it more generic, capturing the real essence of the technology behind, since a DLT could be a DAG (Directed Acyclic Graph) or could be a sequential chain of blocks, or many other implementations. As a bonus, DLT helped me being more detached from the mere financial connotation, improperly associated with blockchain and bitcoin, allowing me and my partners to focus solely on the technological aspect.

"But why don't you just use a database?"; this is probably the biggest comment we've all heard. To me it was also my biggest "reality check" when investigating a new use case. I would ask myself: could this be solved in a simpler way? I've always been a fan of Occam's razor.

Yes, a ledger looks like a database, but it is crucial, though, to understand the key principles behind our ledger:

• Immutability, data become truly trustworthy, hard to counterfeit, highly resistant to attacks
• Shared, or decentralized, in principle all nodes participating in the network hold the same copy of the data

DLT & Industry4.0

As already mentioned, I spent almost 4 years making sense of this technology within the boundaries of the Industry4.0 framework. Unfortunately, I was only able to produce a couple of Proof-of-Concept implementations, but I am still convinced that some of the use cases I collected still have great potential.

When one thinks of Industry4.0 many things come to mind: IoT, AI, edge computing, smart machines, digital twin. I am willing to bet that DLT is not a common one. I quickly came to the conclusion that, unlike all the other listed technologies, DLT have less of an impact of actual day-to-day operations, but rather a much bigger impact as an enabler for new business models. The decentralization and immutability of data allow for new types of interaction with players in the industry, allowing for transparent and trustworthy exchange of data, opening the doors for innovative ways of doing business.

Here are some of the main use cases taken from a long collection I came up with during my investigations, in no particular order:

• Machine-as-a-Service, probably one of the biggest business model innovation made possible by DLT. Just like leasing a car, the idea here is to acquire the "use of an industrial machinery" without actually owning it, hence eliminating the burden of maintenance costs or high initial CAPEX.

• Plant-as-a-Service or Manufacturing as-a-service, similar proposition to the previous one, this time extended to a full set of flexible machineries within a production site. Small manufacturers could be able to prototype with some highly specific (and expensive) equipment just for a small period of time or for a certain amount of parts. Companies could use this concept to test out new technologies before disruption their current processes. Providers could use it to better showcase their new products.
• Data-as-a-Service, concept that has been already put to test by other startups, it will allow data scientist to access certain portion of manufacturing data for a given amount of time for the purpose of training algorithms, data that is usually hard to get ahold of.
• Product Traceability and Provenance, allowing final users to trace quality of good every step of the way. Traceability of raw material and goods has been increasingly demanded over the last decade. A large distributor some years ago was one of the very first to try this for their produce (in that case was mangoes). Another example we saw early was E.coli outbreaks in lettuce, forcing mass distribution to trash entire lots of potentially good products only because nobody could precisely trace the provenance of each lot. DLT could fix this. Similar approach can be used for the automotive aftersales market, to guarantee, for example, that a spare part is original, that a car odometer was not counterfeited, etc.
• Trade Skills and Certification Repository, allowing more transparency and trustworthiness in the skilled labor marketplace. Imagine a repository of certifications (owned by no one organization) where workers can append their achievements and at the same time companies hire only those fit for the job and reduce training efforts.
• Performance-bases Repair Agreement, in the are of service, 35-40% of every OEM revenues come from repair claims. The dark side of this is that it is always hard to identify responsibilities when a machine breaks down: was it misused by the end user (service claim, paid by the user), or was it simply a random fault (warranty claim, paid by OEM)?

Usually, it takes 2-3 weeks to identify the root cause and a team of people behind. What if the machine could tell immediately and without any doubt, what were the parameters when the breakdown occurred. Imagine the machine recording the whole parameter set to a distributed ledger every time a parameter is changed, there will be no more guessing game on whose responsibility it was if a machine spun too fast or wrong current settings were used.

• Digital Bill of Lading, for a faster and smarter supply chain tracking. This is another use case that I investigated thoroughly. Today, all sorts of goods travel the world in large containers using pieces of papers as their bill of lading, a fundamental document stating its content, its value, and other customs relevant data. Custom clearing is still a lengthy and time-consuming activity, accounting for delay in shipments and unhappy customers. What if we could use a trustworthy, immutable, shared data ledger that port and custom authorities can use to be informed of the content of a whole cargo ship, with no possibility of counterfeiting nor loss of paperwork.

Final Thoughts

So, is DLT a viable technology for solving relevant problems in manufacturing? Definitely. It is also true, though, that bringing all the relevant stakeholders at the same table might be a bit more challenging that simply trying out the latest IoT sensor. This time around, players like insurance firms, investing firms, financial banks, but also regulators and government entities could be the driving forces for widespread adoption of DLT-based solutions in an industrial setting.

Looking back, I can identify a few hurdles:

• Skepticism in sharing data. Not all manufacturers are willing to share their manufacturing data, on a shared decentralized ledger. This is true not only for competitors, but even in a direct-to-supplier relationship. The issue here is not really fear for IP losses, but rather unwillingness to share live production data and performance.
• Technological limitations. Depending on the final implementation, a given consensus protocol might create a significant bottleneck in communication. Considering that most architectures are becoming more hybrid cloud-edge, latency can play a decisive role when selecting a technology. Certain projects, using DAG or using virtual gossip, promise to be more efficient and IIoT-friendly.
• Ability to make money. This is a deep philosophical point. We have discussed already how DLT, broadly speaking, is an enabler for new business models. This holds true for the users and players that benefit from transactions of goods, services or data. Not so much for technology providers. The ethos of a distributed ledger is rooted in the idea that everyone must be able to transact with everyone else with no middleman (permissionless), and that is possible because the ledger is shared and replicated in every node of the network. Because of this, a company creating a new distributed ledger implementation is less likely to own, run and operate a massive network of servers which can charge for the service to its users. Obviously there are more centralized permissioned projects out there but the main pain points for a technology provider to create a sustainable business remain.