Blockchain in supply chain management
We argued earlier that supply chain management is one of the industries where blockchain technology is absolutely necessary.
Many participants are involved in the process of distribution and selling goods, be it a piece of clothing or fruit. It involves plenty of paperwork while products or materials pass from one company to another, including manufacturers, warehouses, distributors, retailers, finance companies, etc.
Implementation of blockchain technology into the process can cut expenses significantly and make the distribution process much simpler and more effective. Once the information about goods is automatically recorded on blockchain in a transparent and immutable manner, it becomes accessible by everyone. Thus, every participant of the logistics chain, including the end customer, can check a reliable info about products by simply accessing the records.
How Ambrosus platform works
Ambrosus platform, or Ambrosus network (AMB-NET), is a decentralised blockchain-based data-storage system that helps enterprises improve their supply chain management processes. AMB-NET is a self-sufficient and self-sustainable network, which implements different kinds of masternodes. The consensus is formed using proof-of-authority mechanism. Ambrosus network implements automatic execution of smart contracts, economic incentives for participants and governance procedures.
The data on AMB-NET can be stored either privately or publicly. An example of private data can be the price a ketchup producer paid for the delivery of a container with tomatoes, whereas the ingredients list and expiration date on a yoghurt pack constitute public data. If needed, private information can be shared by its owner.
Among other blockchain platforms that offer solutions for supply chain management Ambrosus is the one that, to a great extent, focuses on quality assurance of products, specifically food and medicine. This assurance is guaranteed via sensors, attached to the product during its lifecycle.
“We go deeper into the linkages between blockchain and sensors,” said Angel. “And provide specific API, through which sensors can connect and pass data to our blockchain. The data can vary from chemical composition of the product to temperature, humidity, brightness or light, GPS location, etc.”
Use case examples
Angel also shared an example of Ambrosus platform use case.
“For instance, if you have a shipment of pharmaceuticals going from the factory through the distribution chain all way to the end user, perhaps you want to know in what locations the product was, the humidity or temperature there, and the owner at a particular time. All this info, gathered with help of sensors and recorded on Ambrosus platform, gives more insights into the functioning of the supply chain than you normally have. You can track the product in real time through the whole way from the point of dispatch to delivery.”
“In technical terms it looks like this. Once a product is produced at the factory and the sensor attached to it, the so-called ‘asset’, tied to the product, is created in a database which is linked to the Ambrosus blockchain. An asset is a digital representation of a real-life object, generated as a numeric/alphabetical string of characters,” added Angel. “As the product travels through the supply chain, different events are recorded and assigned to this particular asset. Those can be readings from sensor, transfer of ownership, or any other events. All of them are recorded in the history of the product. So, the end consumer, whether it’s a pharmacy or actual person buing the product, can access the whole product life history.”
“This data can be integrated into different tools and applications for the supply chain management, such as ERP (enterprise resource planning) software, etc. By order of one German quality assurance company we’ve created a mobile solution which checks vital information about different stages of meat production. And we have also developed a solution where quality control inspectors can use a mobile app to input relevant information about the product for end consumers.”
Data validity and reliability
Another question was how costumers can be ensured that the data recorded on blockchain is correct.
“This is a fundamental question of the interlinkage between a physical object and its digital representation that has been overloked in blockchain in general. On our side we solve it by encouraging producers to integrate sensors into a box or container, and to create different sorts of seals connected to the sensor. Once you open the box or container, the seal becomes broken, the signal is sent to the blockchain that someone opened that package, or that integrity is not maintained. So you should not trust any data since that moment. Smart contracts are watching for these events and as soon as that is recorded, they immediately flag an issue or open an insurance claim, or initiate money reimbursement to client.”
“Another factor is assuring that the integrity and working conditions of sensors have been maintained. So when the sensor sends an information about the product, it also provides a signature that confirms that the sensor is neither hacked nor broken. So you can be sure that only correct data goes onto the blockchain.”
Sensor agnostic platform
Half of Ambrosus team consists of members with microelectronic background, so they could design their own sensors, but they decided to take a different path. Instead, they are aiming to create a protocol that is able to interface with many different kinds of sensors, so anyone could use Ambrosus platform.
“The core value proposition of Ambrosus is to be a sensor agnostic plaform,” said Angel. “In general, we provide data repositories and management systems, and enable as many sensors as possible to talk to the blockchain, but we are not concerned with the way they function. At the same time, members of Ambrosus team can intelligently examine the sensors to recommend which ones are better. For instance, when we run pilots with a large company, they very often have their own sensors in place already. In other cases companies want us to consult them on what technology they should use on a hardware side.”
Also, Ambrosus has R&D lab in Switzerland which develops secure chips for sensors that could be used in next-generation supply chains.
Masternodes and types of data in Ambrosus networrk
There are three different kinds of masternodes within Ambrosus network.
Hermes nodes collect and store data from different sensors, which are connected to AMB-NET via API. After collecting data, Hermes nodes store it locally and publish only its hash to the AMB-NET blockchain.
Normally, owners of Hermes nodes are large companies, which use Ambrosus network in their supply chain management process. They are the original producers and owners of data. They determine which part of data will be private or public, and set expiration date for the data.
You need to have 150,000 AMB stake to run such a node. This stake assures the good behaviour of data producers and discourages spamming the network or other kinds of malicious attacks.
Apollo nodes, which are essentially validators, maintain governance and integrity of the network, they are responsible for execution of smart contracts and processing transactions. Apollo nodes broadcast metadata in the form of hash, provided by Hermes nodes, to the blockchain.
These are the most senior masternodes, since the requirement is 250,000 AMB to run one. These masternodes are rewarded with a transaction fee in AMB tokens for their services.
Atlas nodes store public data. They can also hold a copy of all data stored by Hermes nodes. If there is any issue or an incident where the real data needs to be opened, the public arbiters, represented by Atlas nodes, will always know what the real data was.
Also, Atlas nodes promote information resilience throughout the network by constantly challenging the availability and accuracy of data hosted by Hermes nodes. Thus, Atlas nodes check that the data fingerprint (basically the hash) on the blockchain corresponds to the data that Hermes nodes store locally. If they find out that the data is not there, they can claim the reward from the Hermes nodes in the form of AMB tokens. So, Atlas nodes are constantly supervising honesty and integrity of the network.
You’d need to have at least 10,000 AMB stake to run an Atlas node.
And, finally, there are many Perseus nodes, which are not masternodes. It can be any device interfacting with Ambrosus network. It can be a customer scanning a product bar-code with mobile phone to see the history of the product. It can be a supply chain intermediary recording a particular transaction via readings from sensor.
Angel provided some insights about Hermes and Atlas masternodes.
“We are working on zero-knowledge proof, whereby Atlas nodes would not need to own private data in order to be able to verify the correctness of data of Hermes nodes,” explained Angel. “Moreover, the validation would take place with respect to the contents of bundles of data, not the ultimately private data.”
“Essentially, there are 3 kinds of data in terms of privacy:
Private data (not intended for opening) – only hashes of that data are stored on the blockchain. Private data is always private, contains a lot of sensitive information and would only be required to be revealed by strong push, e.g. in case of public recall, governmental pressure or civil resistance by people. Ultimately, only the company could have the power to reveal that data, but not revealing the data would create very strong negative message to its consumers and partners, letting them know the company is hiding something.
Semi-private data – the actual contents of bundles with detailed insights into assets, events and associated data. Normally, those are stored by data producers themselves, but they can be challenged to be revealed by the Atlas nodes at any moment in time, at which point Atlas nodes get a copy of that data and become data carriers themselves (and must have that data available for public enquiries at any moment).
Public data – is actually the data that is always publicly accessible (akin to data stored on websites). However, in fashion of P2P networks , there are multiple mirrors of that public data, which can be stored on-chain / off-chain and it ensures that public data is always available and has not been altered.”
The main blockchain stores data hashes and verifies that the data hasn’t been changed. But the data itself should be stored somewhere as well. So, in addition to the ability to store data on Atlas nodes, Ambrosus is also developing a solution for distributed storages, although it is in a very early stage. Otherwise, Atlas nodes can become very heavy-weight.
To prevent decentralized storage from overloading, an intelligent solution for removing excess data will be implemented.
In the case of food, it’s definitely possible to remove data in 3 or 4 years, because no one cares about contents of a steak that someone ate several years ago.
But for an aircraft manufacturer, for example, it’s important that the data about the origin of airplane parts be stored for 25 years.
Both short-term and long-term data storage will be possible either on AMB-NET distributed storage, or on company’s local database with fingerprint in the form of hash to be recorded on the blockchain.
Current state of AMB-NET development and transaction speed
The Ambrosus main net is not live yet, but a few test nets are already operating. The biggest test net has about 22-23 nodes.
Main net will go live at the end of July 2018. At the same time two processes will be initiated:
– migration of AMB tokens from ERC-20 to Native Amber on AMB-NET;
– onboarding of masternodes owners/operators, which will initially be given out to the longest-term holders of AMB.
I was curious about how many transactions per second Ambrosus blockchain can handle.
“This is what we are actively testing. Right now it’s around 75-80 transactions per second. It’s a usual performance for proof-of-authority framework,” Angel said.
“The data is put into what we call ‘bundles’. A bundle is basically 10,000 lines in a database. Each line can be either a real life object represented by digital asset on the blockchain, or it can be an event recorded by sensor. Once the bundle is formed, it is dispatched to the network as one transaction,” clarified Angel. “So, we need to understand how many sensor-generated readings can be processed per second, because not all of them have to be transactions. And this is the whole notion of the bundle, because if you apply the notion of the bundle, then it’s 10,000 sensors/events multiplied by 70-80 transactions per second, which gives 700,000 sensor readings per second. That’s an important benchmark.”
Ambrosus is conducting pilot projects with a number of companies. There are about 25 large companies and 50-55 small/medium-sized ones working with Ambrosus in different capacities.
Among major partnerships is the one with Swiss Coffee Alliance, which provides technological and business expertise to companies involved in coffee production and distribution. The other one is with Swiss Food Research, the leading network in Switzerland that connects scientific and innovation projects to the largest agricultural supply chains in the country.
The future of Ambrosus
In the end I asked Angel if Ambrosus platform can be used solely for supply chain management. Or, maybe, it can be implemented in other industries?
“Our current mission is to become a backbone for the smartcities of the future,” he said.” Because we build a protocol which will allow communications between sensors and blockchain. It can be used in quite a wide range of IoT applications. But this is a monumental task, and I’m sure that even in 5 years there will be a lot of problems to solve. So this is our big mission we are working on.”