Blockchain vs. a database: What’s the difference?

Blockchain distributed ledger technology (DLT) has been touted as the answer for just about every transactional issue facing the world today – from payment processing and supply chain tracking to digital identities and copyright protection.

Databases, however, have been serving those same use cases for decades. They record how much money is in a bank account, when cargo reaches a destination and they store the identities of business users – enabling access to business applications and sensitive data.

Because of those similarities, there are cynics (some may even call them pragmatists) who believe once you strip away the hype associated with blockchain and its cryptocurrency origins, what you have left is nothing more than a fancy, but slow and expensive, database.

The argument goes that many of the purported attributes of blockchain can be accomplished with conventional, tried-and-true technology. For instance, there are already hashing algorithms, digital signatures and public key infrastructure (PKI) available for use. If you need a traceable, verified audit trail, you can save your transactions to a database and then digitally sign the data, hash it and store that hash.

The difference: blockchain has all of those features in one place and it plays well with others.

“There is value in blockchain in of itself as a distributed, independently verifiable single version of the truth shared amongst multiple entities – where no one entity is in control and all entities have equal access and equal control,” said Avivah Litan, a Gartner vice president of research.

“It’s also true that you can get non-blockchain technology to support basically the same thing – a distributed independently verifiable single version of the truth shared amongst multiple entities. But, those functions are not built into the technology as it is with blockchain DLT,” Litan added.

The difference between blockchain and a database

At a high level, both traditional databases and blockchain are data storage and data management infrastructures, explained Frank Xiong, Oracle’s group vice president of blockchain product development.

Xiong agreed a traditional database can achieve what blockchain can “technically” by the party that owns and has access to the database. However, if multiple business parties need to perform transactions, they may not necessarily trust a single owner of the database, who creates, updates and keeps all the records.

“The biggest difference is the distributed ledger. We do have distributed databases, but most of them are owned by individual enterprises… where they have databases distributed for different purposes,” Xiong said.

“Blockchain… technology is the preferred technology to create immutable transaction records and keep them in the distributed ledgers, of which every party on the chain has an identical copy and can [have] access to it,” Xiong continued. “In the meantime, it accomplishes immutability, security, privacy and audit capabilities for every party on the chain”

IBM’s vice president of blockchain technologies, Jerry Cuomo, said it helps to visualize databases as birds, with blockchain being a type of bird since it does have the same “DNA”. (IBM is among a large number of software and services vendors, including Microsoft, Oracle, SAP and Amazon Web Services, who offer blockchain-as-a-service to their customers.)

Speaking at IBM’s Think conference in San Francisco last month, Cuomo described DLT as akin to a database but with unique features not exhibited by other types of “birds,” i.e., databases. For example, unlike databases, blockchains have shared ledgers, consensus algorithms, smart contracts and native data immutability –  they are write-once, append many electronic ledgers.

Unlike a database administrator, who has access to commands such as “update” and “delete” that can change records in the ledger, once a transaction has been committed to a blockchain network, its community of administrators are powerless to change it. Each block (or record) is cryptographically secured to the previous block on the ledger, which creates a perfect audit trail.

“Unlike the database bird that has a single administrator who sets up the rules for the ledger, a blockchain… has multiple administrators, each with an exact copy of the ledger,” Cuomo said.

In a database, the administrator controls what data is shared among users, and when a transaction gets submitted, it’s immediately committed to that ledger.

Blockchain DLT is based on a peer-to-peer (P2P) decentralized architecture, with  multiple administrators as part of its consensus protocol. In other words, transactions on a blockchain network are first proposed, then consented to by the group. Only if 51% of that group agree the transaction is acceptable is it then added to the ledger.

Blockchain’s consensus protocol also means it is fault tolerant and it can continue to operate even in the presence of bad actors because a majority of users will keep transactions honest.

Permissioned vs. public blockchains

Not all blockchains are alike. For example, some blockchains are public – such as bitcoin – while others are private or permissioned, such as Hyperledger Fabric, R3 Corda and Ripple. In a public blockchain, anyone can sign up to become another node in the network and submit transactions to it. And, anyone can see those records (such as bitcoin transactions).

In a permissioned blockchain, the originator of the ledger determines who can join, see the transactions and submit new blocks. Yet, every authorized node in the chain still has say over what data is approved for the record. Network members are known and identified by membership PKI keys issued by a decentralized certificate authority.

Additionally, the promise of decentralized consensus on top of permissioned blockchain transactions could eventually enable parties anywhere who don’t necessarily trust each other to conduct business in a trusted fashion, according to Litan.

Unlike a database, theoretically, each entity that participates in a permissioned blockchain network can run a consensus/validation node; in practice, they don’t, because they don’t have the skills or the bandwidth for it. Instead, they typically relegate it to the project sponsor or vendor, Litan said.

“The common thinking out there is that once these companies get comfortable and gain expertise with blockchain, they will participate in transaction validation and consensus, along with the project sponsor or vendor,” Litan said. “However, that isn’t happening in the immediate future – not until public blockchain matures and scales.”

Permissioned blockchains offer business automation tools through smart contracts. Smart contracts execute transparent, pre-determined rules and enable blockchain to avoid a central authority. Smart contracts operate on an “if this happens, then this executes” premise. For example, when a shipping company receives payment for a delivery, a smart contract in a supply chain blockchain can trigger the supplier to manufacture another product to fill the next order.

One myth is that once a smart contract has been written, a mistake can’t be fixed or changes can’t be made. In other words, you’re stuck with the bad code.

Not so.

“In a permissioned environment, the ability to update smart contracts is a given, and it’s designed into the frameworks,” said Martha Bennett, a Forrester principal analyst. “You, of course, need a strong governance model, but then you also need that for public [blockchains] – technically, it doesn’t take much to fork a chain.”

Governance models allow blockchains to temporarily or permanently split or “fork,” creating a new branch of blocks. A hard fork is a permanent divergence from a previous blockchain; a soft fork is a temporary change that is backward compatible. Think of a railroad train changing tracks through a switch; in a blockchain, the switch would be governed by the majority with power over the railroad service.

For example, the Linux Foundation’s Hyperledger Fabric is a permissioned blockchain platform. That means all participants are, to an extent, identified and that the blockchain comes with a proper governance to resolve issues that may arise.

When to use a blockchain instead of a database

Forrester’s Bennett said companies shouldn’t really use a blockchain unless the use case really calls for that type of architecture, because distributed systems always add overhead, and many of the algorithms and techniques are still in their infancy.

There are two key questions for a company considering blockchain:

1. Does the ecosystem (or the initiator of the distributed ledger network) have a good reason for not wanting to share data via a single, centrally controlled system?
2. Does the company want to address use cases involving automated processes running across corporate boundaries and/or leverage the potential of tokenization? (Tokenization is the digital representation of a commodity, such as money or physical goods.)

Where blockchain technology shines is when multiple organizations are involved, according to Joel Weight, COO at Medici Ventures. (Medici Ventures, the venture capital arm of Overstock.com, has been investing heavily in blockchain technology, including dozens of start-ups. Five years ago, Overstock.com began accepting bitcoin as a form of payment.)

“My bank doesn’t need a blockchain to track the balance of my checking account, or to transfer funds from my checking account to my savings account,” Weight said. “In that case, the bank is just moving money from one pocket to another, which a fast, secure database is perfectly suited for.”

Where blockchain can be useful is when two organizations have a proprietary view of the world, each stored in their own databases. In order to share that data, there’s a cost involved to ensure each company’s view of the data is the same or to ensure both parties actually have the assets they expect to exchange, Weight explained.

For example, if instead of using escrow services or following an expensive, slow protocol, all parties work from the same data, then costs to normalize data and trust is minimized. Attepting to do that with a database would require one company to be the owner of all the data – and the source of “truth” for everyone involved in the transcation.

“The permissioned blockchain only requires enough trust between institutions to decide who will participate in the network, then the blockchain takes care of keeping participants honest once they join the network,” Weight said. “The permissioned blockchain network would allow the atomic transfer of value between institutions where both can instantly agree on the state of the ledger before and after the transaction has occurred.”

IDC

Though many companies still confuse the differences between blockchain and traditional databases, DLT seems set to grow quickly over the next few years as companies move beyond pilot programs.

Worldwide spending on blockchain solutions is forecast to be nearly $2.9 billion this year, an increase of 88.7% from the $1.5 billion spent in 2018, according to a newly updated Worldwide Semiannual Blockchain Spending Guide from IDC.

“Blockchain is maturing rapidly, and we have reached an inflection point where implementations are moving quickly beyond the pilot and proof of concept phase,” said James Wester, research director for IDC’s Worldwide Blockchain Strategies.