Node
Any computer that participates in a blockchain network by storing and propagating data. Nodes can be full (verifying everything), light (verifying headers), or specialized (e.g., archive, validator).
What nodes do
Different node types have different roles:
- Full nodes — verify everything; maintain complete state and history.
- Light nodes — verify subset; rely on full nodes for some data.
- Archive nodes — full state plus historical snapshots; useful for analytics.
- Pruning nodes — full nodes that drop old state to save disk.
- Validator / miner nodes — produce or attest to new blocks.
- RPC nodes — serve data to applications and users.
A blockchain network is the collective set of all these nodes operating together.
What network nodes provide
The functions:
- Verification. Each full node enforces protocol rules; collectively they define what the chain "is."
- Censorship resistance. Distributed nodes make individual censorship impossible.
- Data availability. Multiple nodes ensure historical data remains accessible.
- Connectivity. Nodes propagate transactions and blocks throughout the network.
- Service provision. RPC nodes serve data to wallets, dApps, and indexers.
Bitcoin and Ethereum node counts
Approximate scale:
- Bitcoin — ~15,000-20,000 reachable full nodes globally.
- Ethereum — ~5,000-10,000 reachable full nodes.
- Validator counts — Ethereum has ~1M+ validators; Bitcoin doesn't have validators (uses miners).
These numbers are healthy from a decentralization perspective but somewhat concentrated in cloud infrastructure.
Running a node
Resource requirements vary:
- Bitcoin full node — 600+ GB SSD, 4 GB RAM minimum, reliable internet.
- Ethereum full node — 1+ TB SSD (NVMe recommended), 16 GB RAM, reliable internet.
- Solana validator — multi-TB SSD, many cores, fast network.
- Various other chains — varying requirements.
For Bitcoin, individuals can run nodes practically. For higher-throughput chains, requirements increase substantially.
Why running a node matters
Several practical benefits:
- Self-verification. Don't trust third-party RPC services; verify yourself.
- Privacy. Your queries don't leak to third parties.
- Network contribution. Your node helps the network's overall health.
- Reliability. Not dependent on any single RPC provider.
RPC providers
Most users don't run nodes directly:
- Alchemy, Infura, QuickNode — major Ethereum RPC services.
- Chain-specific RPC services for various chains.
- Built-in wallet RPCs — wallets often integrate RPC services automatically.
These provide convenience but introduce dependency. If they fail or get compromised, applications relying on them fail.
Node centralization concerns
A persistent worry:
- Many full nodes run on cloud providers (AWS, Hetzner, others).
- Cloud-provider concentration creates potential single points of failure.
- Geographic distribution is real but limited.
- RPC service concentration affects how dApps access chains.
These don't directly compromise network security but produce subtle centralization risks.
In broader infrastructure
Nodes are one layer in deeper stack:
- Mining/validation hardware at the bottom for produced chains.
- Full nodes for verification.
- Indexers and analytics consuming chain data.
- APIs and RPC services abstracting node operation.
- Wallets and dApps consuming services.
Each layer can be more or less centralized; node decentralization is one component of overall chain decentralization.
What individuals should know
For most users:
- You don't need to run a node for most use cases.
- Wallet apps handle complexity through RPC services.
- Running a node adds privacy and verification benefits but operational cost.
For high-value holders or privacy-sensitive users:
- Running your own Bitcoin node is practical and provides meaningful benefits.
- Ethereum nodes require more resources but provide similar advantages.
- Specialized tools (Umbrel, MyNode) make running nodes accessible without deep technical expertise.
For developers:
- Reliable RPC access is critical infrastructure.
- Backup RPC providers reduce single-point-of-failure risk.
- Self-hosted nodes for production-critical applications.
The basic principle: nodes are the network. Even users who don't run them benefit from their collective operation. Running one when feasible adds personal benefits and supports the network's decentralization.