What a blockchain actually is
A blockchain is a record book that many computers keep copies of and that no single party controls. Entries are collected into groups called blocks, each block is tied to the one before it, and the network follows shared rules for deciding which new blocks are accepted. The result is a history that is hard to change quietly after the fact.
That description is deliberately plain. Most of the confusion around the subject comes from mixing up the underlying data structure with the things people have built on top of it, such as currencies, collectibles and speculation. This article is about the data structure and the problem it was designed to solve.
The problem it was built for
Imagine a group of people who want to share a record, such as who owns what, but who do not fully trust one another and do not want to rely on a single referee. Normally a bank, a registry or a company would hold the master copy, and everyone would trust that keeper to be honest and competent.
A shared ledger tries to remove the need for that single keeper. Everyone holds a copy, and rules built into the software decide how new entries are added. If one participant tries to rewrite history, the other copies disagree and the attempt is visible.
The main parts
Several ideas work together, and it helps to take them one at a time.
- Blocks and links. Entries are bundled into blocks. Each block includes a short digital fingerprint of the previous block, which is the job of cryptographic hashing. Change an old entry and its fingerprint changes, which breaks every link after it.
- Nodes. These are the computers that keep copies and check new entries against the rules. Anyone running the software can usually inspect the record, depending on the design.
- Agreement rules. Because there is no referee, the network needs a method for choosing which new block counts. These methods are called consensus mechanisms, and they differ widely in cost and in what they assume about participants.
- Digital signatures. Entries are typically authorized with a pair of keys: a private key that only the owner holds, and a public key that lets others verify the signature. Losing the private key generally means losing access, and there is usually no help desk to call.
Public and private designs
Not every blockchain is open to everyone. A public design lets anyone join, read and propose entries. A permissioned design limits participation to approved members, such as a group of companies or institutions that already know each other.
Permissioned systems give up some of the “no trusted party” idea in exchange for speed, simpler governance and easier compliance with legal rules. Whether that still counts as a blockchain or is better described as a shared database with extra safeguards is a matter of ongoing debate, and reasonable people draw the line in different places.
What “tamper-evident” really means
A common claim is that blockchains are unchangeable. A more careful statement is that they are tamper-evident and costly to alter. Altering old entries would mean redoing the linked fingerprints and persuading enough of the network to accept the rewritten version. In well-run public systems that is expensive or impractical, but it is not a law of nature, and smaller or poorly designed networks have been successfully attacked.
Also, a ledger only records what it is told. If false information goes in, the record will faithfully preserve the falsehood. A blockchain can show that an entry was made and has not been altered since; it cannot show that the entry was true to begin with. This gap between the digital record and the physical world is one reason why applications such as supply chain blockchain projects depend heavily on the people and sensors that feed data in.
Where it can fit
A shared ledger is most plausible when several of these conditions hold together:
- Multiple organizations need to write to the same record.
- They do not want any one of them to own the master copy.
- A history that cannot easily be rewritten has real value.
- The cost and slowness of coordinating many copies is acceptable.
Examples that people have explored include recording transfers of digital assets, tracking the movement of goods between companies, and keeping registries that several parties rely on. Results vary, and many pilot projects have been quietly dropped, so it is wise to treat any specific claim of success with some caution.
Where it usually does not fit
If one organization controls the data and everyone already trusts it, a conventional database is typically faster, cheaper and easier to fix when something goes wrong. Ordinary databases can also be backed up, audited and edited to correct mistakes, which is sometimes a feature rather than a flaw.
Other trade-offs are worth knowing:
- Speed and size. Having many nodes check and store every entry usually makes a blockchain slower and less efficient than a central system.
- Privacy. A record that many people can read can conflict with data privacy expectations and laws, especially around the right to delete personal information. Pseudonymous is not the same as anonymous, and activity linked to an address can sometimes be traced back to a person.
- Mistakes are hard to undo. A wrongly sent transfer or a flawed rule may be impossible to reverse without broad agreement.
- Energy. Some designs use a great deal of electricity and others very little, depending mainly on the agreement method.
Blockchain versus the things built on it
It is useful to keep a few separate questions apart. Whether the underlying ledger technology is sound is one question. Whether a particular token, product or company is a good idea is a different one, and many of the well-known failures involved the second rather than the first. Fraud and speculation happen on top of the technology, and the same technology can be used for dull, useful purposes with no speculation at all.
Because outcomes in this area are uncertain and the field changes quickly, nothing here is a prediction about which uses will last, and none of it is financial advice.
A short way to judge a claim
When someone says a blockchain will fix a problem, a few questions usually clarify matters. Who writes to the record, and why do they not trust each other? What stops false data from going in? Who is responsible if something goes wrong? Would a normal shared database do the same job? If the answers are vague, the blockchain may be adding complexity rather than value.