Blockchain-Based Databases

Blockchain is one of the most promising technological trends in today's financial services landscape. It represents the underlying technology that enables the so-called cryptocurrencies such as Bitcoin. In this article, I will briefly walk through the basic idea behind blockchain and its applicability as a data storage system.

At the most basic level, a blockchain is a data structure that stores data as a chain of linked information blocks. The most prominent application of blockchain technology is in the creation of digital ledgers. A digital ledger is a record-keeping book of accounts that keeps track of a business’s financial transactions. Think for example when you deposit or withdraw money from your bank account. As per common sense, you trust that your bank will maintain a correct and truthful ledger and not modify it by deleting or updating book records.?

A blockchain data ledger performs a similar function and guarantees data immutability, but follows different design principles. To illustrate the idea, let’s represent a blockchain as a linked list of information blocks, as illustrated in the figure below. The first block in a blockchain is called the genesis block. Each subsequent block contains its own data (i.e. transaction records) as well as a hash pointer that stores both a pointer to the previous block as well as a hash of that older block. A hashing function H() is used to create a cryptographic hash of each block’s data. A block hash serves as a unique identifier of each block.

To illustrate why this structure is tamper-resistant, consider an adversary attempting to manipulate the blockchain by altering data entries in block 2. This will automatically update the block's hash. As a result, the hash pointer in block 3 will no longer correspond to the updated hash of block 2. To conduct a successful attack, the adversary must alter all hash pointers up to the most recent block in the list (i.e., the head).

In addition to data immutability, blockchain-based data ledgers have demonstrated remarkable reliability when implemented as a distributed system. The term Distributed Ledge Technology (DLT) is often used to describe such systems. In a DLT, a decentralized network of nodes cooperatively maintains the integrity of the ledger. Each node keeps a copy of the blockchain and can verify and validate any operation that alters the blockchain. For example, if a node wants to add a new block to the chain, the other nodes in the network need to agree and reach a consensus on such an operation. The same applies to our earlier example of the adversary attack: even if the adversary were able to change all the hash pointers, the changes would still need to be validated and accepted by the other nodes in the network. A variety of consensus mechanisms are available, including Proof of Work, Proof of Stake, and Byzantine consensus.

At this point, you might be wondering if a blockchain is a good solution as a data storage system. Technically speaking, a blockchain can be used to store transaction data, but it comes with a price. First, blockchains have very limited data querying capabilities. Second, as more nodes are added to the network, performance may get worse in terms of throughput, latency, or capacity.

To overcome these issues, blockchain-based databases (or blockchain databases for short) were introduced. The main idea behind a blockchain database is to combine the best of both worlds: the performance and power of a traditional database system (e.g. SQL or document databases) with the immutability, decentralized control, and traceability of a blockchain.

Crucially, when deployed internally within a firm, a blockchain database doesn’t need to be decentralized. The firm in this case acts as the administrative authority that controls the blockchain.?

There exists already a few commercial solutions for blockchain databases. The most prominent is BigchainDB, which uses MongoDB as the distributed database under the hood and offers blockchain characteristics, and Amazon Quantum Ledger Database (QLDP), a fully managed ledger database for creating an immutable, transparent, and cryptographically verifiable transaction log.

#blockchain #blockchaindatabase