Secure Document Transfer Built on Top of Blockchain Technologies

According to the survey results, the respondents believe that in 2016 the blockchain trend is likely to accelerate, resulting in new promising partnerships and high-tech product launches that would bring in major investments.

And it looks like the prediction has already started becoming a reality. According to cryptocoinsnews.com, Microsoft has recently announced its collaboration with Bank of America, with the aim to “fuel the transformation of trade finance transacting, with blockchain technology.”

The other important outcome of Deloitte's survey regards the emerging new use cases of the blockchain and the decentralized approach itself, as this technology seems to have a lot of potential to reach beyond the financials, spreading into real estate, commodities, retail, digital media etc. This is just a short list of domains where a lot of experimentation with blockchain has started recently. And we at ELEKS also joined the trend with the number of experiments, building our own custom applications on top of the blockchain technologies. Below are the stories of two such experiments.

The first one was a simple experiment with parsing the Bitcoin blockchain to find "the richest" wallet. You can read about this experiment below.

However, the most exciting part of our experimenting with Blockchain is a solution for secure document transfer. This article describes how we started with internal PoC and ended up with a real prototype for one of our clients from the finance industry. You’ll see how we advanced our approach and how the technology stack evolved.

When we developed the secure document transfer system for our customer, we opted for blockchain for some specific reasons — this technology allows to establish an environment that combines efficiency, security, transparency and trust among business participants.

In general, the speed of blockchain transactions may be lower than it is with databases. This is because, with blockchain, the system needs to perform some extra actions such as verifying the consensus and applying cryptography to ensure the safety of transactions. However, thanks to these additional steps, blockchain provides trust and security, even in semi-trusted environments, since the transaction history is available to every stakeholder instantly.

“The richest wallet”: our first experiment with blockchain

Our experiments with this technology started in 2015 with the attempts to parse the Bitcoin blockchain, looking for the richest bitcoin address or “wallet”.

Technically, a wallet is an address key that contains private keys for certain Bitcoin addresses that belong to the user who owns the balance. At that time, the difficulty in reaching the project’s goal was that we had to download large volumes of data and store them locally. That data had to be parsed and all information about the user balance had to be stored in memory, so we could process and analyze it. As a result of our first experiment with blockchain, we managed to build several graphs.

The below graph represents the total number of transactions taking place over the last 5 years.

Having analyzed this data, we organized it in regard to the transactions split by days over the period of the last 180 days. Here is the result.

After we had all the data organized in this way, we managed to define the top addresses, in other words - the wallets with the highest balance.

With the help of this experiment, we mainly aimed to explore the Bitcoin blockchain. To get more hands-on experience with this technology, we created our own Bitcoin blockchain parser with the help of the .Net Framework. As soon as we learned how the blockchain works and how it is organized from the inside, as a next step, we decided to build our own custom software on top of the blockchain.

We are passionate about new technologies and exploring their potential value for our customers’ businesses. As blockchain is an extremely popular and promising trend now and it spreads quickly across industries, we tested this technology to be able to quickly implement our customers’ ideas through R&D experiments and PoCs created on top of the blockchain infrastructure.

Tapping into the capability of smart contracts

While we started experimenting with blockchain technologies, we learned that the decentralized approach actually allows one to organize a secure transfer for various types of data and its potential is not limited to financial transactions. And, with the introduction of the smart contracts, this capability has been extended significantly.

Smart contracts are self-automated computer programs that can self-execute and self-enforce by using some pre-programmed conditions and run without any possibility of downtime, censorship, fraud or third-party interference.

Designed to be very similar to their paper equivalents, smart contracts can carry out the terms of any kind of agreement. They feature the two parties involved (the sender and the receiver), include the reference to the document that is being transferred as well as the other attributes, such as the timestamp, block-address, self-address etc.

The key aim of this technology was to streamline financial transactions, allowing for business operations to be done by two anonymous parties through the Internet with no middleman required. However, smart contracts are not limited to finance. As this technology matures, it allows for storing and transferring the assets referring to the legal sphere, real estate, intellectual property and more. Inspired by this broad tech potential, we decided to create a custom system for smart document management on top of the blockchain, allowing one to securely store and transfer various kinds of assets.

In other words, blockchain supports a different aim comparing to databases. Blockchain uses a decentralised approach, while with a database, users have limited access and the administrator can modify the data at any time.

Developing a system for secure document transfer powered by Ethereum

We decided to build our system with the help of Ethereum, a decentralized platform that runs smart contracts.

One of the reasons we opted for Ethereum is because it supports a wide range of possibilities for smart contract development. This platform also provides an advanced programming language with a large variety of data types, functions and so on. Another benefit of Ethereum is that this network is live; it already works. All these things together made Ethereum an optimal choice for us to experiment with blockchain.

Ethereum blockchain is a permissionless blockchain, where anyone can join the network as well as write and read transactions. But usually, businesses need private blockchain networks so that only predefined participants can join a network, view and make transactions. In this case, permissioned blockchains should be used (e.g. Hyperledger blockchains).

Just recently (summer 2017) Hyperledger evolved and released a new version with more functionality and tools, which made it an interesting platform for private or semi-private blockchains. Hyperledger suited our client’s business better because it allows establishing more granular control over who sees what and who has access to what, compared to public blockchains — where all information is available to everyone, and we have to think about additional cryptography.

The aim of this experiment was to provide secure storage and transfer for various kinds of financial, legal or any other types of documents, such as agreements, ownership clauses, private photos and more. In other words, we wanted to create an environment where legal transactions can be processed safely and with no third-party intermediary required. In daily life, when people conclude legal agreements, the process requires the involvement of a notary whose job is to verify the agreement, sign it and register in his/her journal. While, in the blockchain world, you do not need to pay for notary services, which usually cost a great deal of money. Thanks to cryptography, all the signatures can be secured while blockchain can do the verification and maintenance of the registry journal.

And more importantly, it takes much less time, especially when the parties are in different time zones. Imagine a logistics organisation in Amsterdam that needs to sign an agreement with the port of Singapore and a bank in Dubai. In case of conducting a regular legal contract, the parties may need to organise video calls and sign the document when all the representatives are available online. With the blockchain solution - all participants can grant their approval whenever they are available; there is no need to coordinate this process — the business logic and smart contracts will do the job.

The solution we were creating included two key units: document storage and a smart contracts journal.

We wanted the storage to contain all the documents, providing public access for verification and private restricted access for reading, and therefore we chose Github for this purpose. Having the files located in public GitHub, we made them visible to everyone, but still they were secured with encryption. Combining public storage and cryptographically-secured access, we made it possible for anyone to access the file; however, without an encryption key, this file was only a plain set of bytes.

In the case of our customer, instead of using a public GitHub we opted for IPFS private network as storage. IPFS, the InterPlanetary File System, is a content-addressed peer-to-peer distributed file system. The stakeholders can upload files into the network, and these files are shared among peers. The participants can upload files to the system with the help of an IPFS client, which is an entry point to the network.

IPFS network can also be accessed through publicly available gateways directly from a web browser, which is a massive advantage to those users that have no IPFS client installed on their devices. To ensure security, IPFS uses encryption for all the transactions made. Private Networks allow IPFS to connect only to other peers who have a shared secret key. All nodes in a private network are isolated from a public network, and the data is shared only between nodes in a private network.

Here's a step-by-step guide for building your own private IPFS network with IPFS-Cluster for data replication.

Our solution encrypts the secret key with the recipient's public key so that only the recipient can decrypt it with his/her private key. The encryption keys were sent through the Ethereum blockchain with the help of smart contracts. These contracts contained the sender’s and receiver’s addresses, a key, as well as a link to the publicly-stored document. The contract rules were preprogrammed in a way that only the receiver could obtain a key.

We used the smart contracts journal to contain all the transactions as well as their attributes. The journal also provided an interface for private access.
We decided to add the option of changes tracking as well. Therefore, the sender and receiver could change the existing documents and all the updates could be tracked via blockchain. This way, everybody would see that there were some changes implemented, but again, no one would be able to check what was changed without an encryption key.

The blockchain network that we created contained a number of nodes: mining nodes and client nodes. Mining nodes contain a full version of the blockchain and perform the “proof-of-work” consensus algorithm to mine new blocks. Client nodes also contain a full blockchain and do not do any work except for providing a blockchain interface to enable smart contracts creation and calls. In our case, we decided to keep the blockchain in the client nodes so we could easily switch the node types, but it is still possible for the client node to keep only parts of the blockchain.

The mining nodes are hosted on Azure as a virtual machine with an Ubuntu operating system. Client nodes are hosted locally (as Geth - a command line interface for the Entereum blockchain) together with the user interface written on .Net WPF. The client WPF application includes all the logic required to work with GitHub as a document storage and Geth as a blockchain interface. And the reason why we opted for WPF is that the WPF application’s interface is far more user-friendly, compared to a “black command line”. The WPF application also provides the logic for account creation and the wallets are stored in local Geth node.

Since the first version of the application had been built with .Net WPF, it was intended for Windows only. While our customer needed a cross-platform solution, we opted for Node.js to make the app run on Windows, Linux and MacOS.

Apart from Node.js, we used C++ and Go for backend development to make the system perform efficiently under high loads. The below scheme shows our network in detail.