Blockchain
Blockchain
Definition: A distributed, append-only ledger where transactions are grouped into blocks, cryptographically linked to the previous block, and replicated across many independent nodes.
How It Works
- Transactions are broadcast to the network, collected by nodes, and grouped into a candidate block along with a timestamp
- Each block header stores a cryptographic hash of the previous block’s header, this hash pointer is the “chain” in blockchain
- Transactions within a block are organized into a Merkle tree, so any single transaction can be verified against the block without downloading the whole block
- Before a block is accepted, the network must agree it’s valid through a consensus mechanism, no single party appends blocks unilaterally
- Full nodes each store a complete copy of the ledger and independently re-check every new block against the same validation rules, they don’t trust another node’s word for it
- Light (SPV) nodes skip storing the full chain and instead verify just block headers plus a Merkle proof for the transactions they care about
- Occasionally two valid blocks are produced at nearly the same time, a fork, and nodes temporarily follow different chains until the network converges on one, typically the chain with the most accumulated work or stake behind it
- New blocks reference exactly one parent, so the ledger’s history forms a tree in principle but converges to a single accepted chain in practice
Network Types
- Public/permissionless: anyone can run a node, validate, or submit transactions, e.g. Bitcoin, Ethereum
- Private/permissioned: only approved participants can validate, used inside a single company or consortium
- Consortium: a fixed group of organizations jointly operates the validating nodes, common in interbank settlement pilots
- Hybrid: a public settlement layer with a permissioned layer on top for specific participants, used by some enterprise chains
Block Anatomy
| Field | Purpose |
|---|---|
| Previous hash | Links this block to its parent, forms the chain |
| Merkle root | Single hash summarizing every transaction in the block |
| Timestamp | When the block was proposed |
| Nonce | A value miners vary to find a valid Proof of Work hash |
| Transaction list | The actual transfers/contract calls included |
Forks
- Soft fork: a backward-compatible rule tightening, old nodes still accept new blocks as valid
- Hard fork: a rule change that old nodes reject, splitting the network into two chains unless everyone upgrades
- Orphan block: a validly mined block that loses the race to another block at the same height and gets discarded by the eventual winning chain
Under the Hood
Hash-linking makes the ledger tamper-evident rather than tamper-proof. Altering history is technically possible, it’s just prohibitively expensive once enough blocks have been built on top of the one you’d need to change.
Hash-Linked Blocks
Block Proposal and Validation
Worked example: why editing one block breaks the chain
- Given: Block 101 originally records “Alice sends Bob 5 coins” and hashes to
0x77e9, Block 102’s header storesPrev: 0x77e9 - Step: an attacker edits Block 101’s content to say “Alice sends Bob 50 coins”
- Step: since a hash is a function of the block’s content, Block 101’s hash changes from
0x77e9to something like0x93bc - Step: Block 102 still stores
Prev: 0x77e9, which no longer matches Block 101’s real hash, so Block 102 fails validation, and every block after it too - Answer: the attacker must recompute every block from 101 to the current tip and get honest nodes to accept that alternate chain, before they extend the real one further
Worked example: Merkle proof for one transaction
- Given: a block contains 2,048 transactions, a wallet wants to confirm its one transaction was included without downloading all 2,048
- Step: the block header stores only one Merkle root hash, built by repeatedly hashing pairs of transactions together up a binary tree
- Step: the wallet requests a Merkle proof, roughly log2(2048) = 11 sibling hashes along the path from its transaction to the root
- Answer: hashing its transaction up through those 11 sibling hashes and comparing the result to the block’s published Merkle root confirms inclusion, without downloading the other 2,047 transactions
Why It Matters
- Enables trustless coordination between parties who don’t know or trust each other, no bank, court, or company has to vouch for the ledger’s accuracy
- Removes a single point of failure, there’s no central server to hack, subpoena, or shut down to alter the record
- Makes the full transaction history independently auditable by anyone running a node, rather than trusted to a private party’s word
- Underpins everything else in this glossary section, smart contracts, NFTs, and DeFi all depend on the base ledger’s guarantees
- Gives users direct custody of assets through a Wallet, instead of a bank holding the actual ledger entry on their behalf
- Provides a common settlement layer that different applications and organizations can build on without negotiating bilateral trust agreements
Common Pitfalls
- Assuming “blockchain” automatically means “secure” or “better,” most applications don’t need a decentralized, trustless ledger, and a normal database is faster, cheaper, and easier to fix when something goes wrong
- Confusing the blockchain (the ledger and its consensus rules) with cryptocurrency (one application built on top of it), plenty of blockchain use cases don’t involve a tradable coin at all
- Treating “on the blockchain” as synonymous with “permanent and true,” a bug or bad input can still permanently record wrong data, immutability preserves whatever was written, not whether it was correct
- Underestimating how few nodes some smaller chains actually run, a chain with a handful of validators is far easier to attack than the “thousands of nodes worldwide” framing implies
- Forgetting that “immutable” is a matter of degree, a 51% attack or a coordinated hard fork, as happened after The DAO hack on Ethereum in 2016, can rewrite recent history in practice
- Overestimating throughput, base-layer blockchains typically process single-digit to low hundreds of transactions per second, far below centralized payment networks
- Confusing confirmation count with finality, a transaction with one confirmation can still be reorganized out on some chains, most wallets and exchanges wait for several before treating it as settled
Comparison
| Public blockchain | Private/permissioned blockchain | Traditional centralized database | |
|---|---|---|---|
| Who can validate | Anyone (Bitcoin, Ethereum) | Approved participants only | N/A, single operator |
| Trust model | Trustless, rules enforced by consensus | Trust the approving organization(s) | Trust the database owner |
| Speed/throughput | Slower, global consensus overhead | Faster, fewer validators | Fastest |
| Censorship resistance | High | Low to medium | None |
| Data can be deleted | No, only appended | Usually no | Yes |
| Governance | On-chain votes or rough consensus among developers | Controlling organization decides | Owner decides |
| Typical use | Public money, open protocols | Interbank settlement, supply chain | Most business applications |
Example
Bitcoin’s blockchain is a public ledger of every transaction since block 0, the 2009 genesis block, replicated across thousands of independently operated nodes worldwide, none of which needs permission to join or verify the chain.
Related Terms
Referenced by