Blockchain

What makes a blockchain different from a regular database

A regular database has one operator who controls writes and can change history. A blockchain replicates the data across many independent operators, requires them to agree on changes through a consensus mechanism, and structures the data so that altering past entries breaks the cryptographic linkage between every subsequent entry. The result: a record that no single party can rewrite and that anyone can independently verify.

The trade-offs are real. Blockchains are slower, more expensive, and less feature-rich than traditional databases. But for use cases where independent verifiability and resistance to centralized control matter — money, ownership records, agreements between parties who don't trust each other — those properties are worth the costs.

How a block becomes part of the chain

The mechanic, in five steps:

1. Users sign and broadcast transactions to the network.
2. Transactions accumulate in a "mempool" of pending operations.
3. A node (miner or validator) bundles transactions into a block and proposes it.
4. Other nodes verify the block satisfies protocol rules: signatures valid, no double spending, block-size and fee constraints met.
5. Once the block accumulates enough acceptance, it's appended to the chain. Each node's local copy updates.

Each new block contains the hash of the previous one. Tampering with any historical block changes its hash, which invalidates every block built on top of it — so changing the past requires redoing all the proof-of-work or proof-of-stake security expenditure since.

Public vs. private vs. permissioned

Three rough categories:

• Public, permissionless — anyone can read, write, and run a node. Bitcoin, Ethereum, Solana. The pure form, with the strongest decentralization properties.
• Public, permissioned — anyone can read, but only approved entities can write or validate. Hybrid model used in some enterprise contexts.
• Private (consortium) — both reads and writes restricted. Used inside organizations or among small groups of cooperating parties. Generally provides little advantage over a regular replicated database, which is why this category has mostly faded from interest after the 2017–2019 enterprise blockchain hype cycle.

The interesting blockchain activity is overwhelmingly on public permissionless chains.

What blockchains do well

A few use cases where blockchains have meaningful advantages over traditional infrastructure:

• Money without trusted issuers — Bitcoin solved the digital-scarcity problem; this is the original and still most important use case.
• Programmable financial primitives — DeFi built lending, exchanges, and derivatives that anyone can use without intermediation.
• Verifiable scarcity for digital assets — NFTs made it possible to own a specific digital item provably.
• Cross-border value transfer — stablecoins on blockchains move billions per day in seconds, at a fraction of correspondent-banking costs.
• Censorship-resistant publishing — text, code, and data stored on or referenced by chains is hard to take down.

What blockchains don't do well

Equally important to be honest about:

• High-throughput general computation — chains are far slower and more expensive than centralized cloud computing for non-trust-critical workloads.
• Storing large data — most chains charge per byte and aren't designed for files; data is typically stored off-chain (e.g., on IPFS) with the chain holding only a hash.
• Privacy by default — most blockchains are transparent. Privacy requires deliberate use of zero-knowledge proofs, privacy coins, or other specialized infrastructure.
• Reversing mistakes — finality means typos and exploits can't be undone by an admin.

The architecture is splitting

The "monolithic" blockchain (one chain handling consensus, execution, settlement, and data availability) is being challenged by modular blockchain designs that split these functions across specialized layers. Layer 2 rollups like Arbitrum and Base handle execution while settling to Ethereum; data availability layers like Celestia store rollup data; chains like Solana and Monad take the opposite bet that monolithic designs can scale through parallelization. The eventual winner is unsettled, and the answer may be "different architectures for different use cases" rather than convergence on one model.

Beyond crypto

For a few years after 2017, "blockchain" was a buzzword applied to almost any distributed-ledger experiment. Most non-crypto blockchain pilots in supply chain, healthcare records, and trade finance produced little durable value because they didn't actually need the trust-minimization properties — a regular shared database with proper governance would have worked. The robust uses of the technology have stayed concentrated in finance, gaming, identity, and digital ownership, where the censorship-resistance and verifiability matter.