Cluster Replicated Data Store—the data structure underlying Solana's gossip protocol. CRDS uses a push/pull model where validators exchange versioned, signed records (contact info, vote state, epoch slots). Records have wallclock timestamps and are pruned when stale. Each record is keyed by (source pubkey, label).
Start with the shortest useful explanation before going deeper.
Cluster Replicated Data Store—the data structure underlying Solana's gossip protocol. CRDS uses a push/pull model where validators exchange versioned, signed records (contact info, vote state, epoch slots). Records have wallclock timestamps and are pruned when stale. Each record is keyed by (source pubkey, label).
Use the quick analogy first so the term is easier to reason about when you meet it in code, docs, or prompts.
Think of it as part of the chain machinery that keeps ordering, execution, or consensus moving.
Place the term inside its Solana layer so the definition is easier to reason about.
Consensus, leader rotation, slots, epochs, and the runtime.
Turn the term from vocabulary into something operational for product and engineering work.
This term unlocks adjacent concepts quickly, so it works best when you treat it as a junction instead of an isolated definition.
Use this compact block when you want to give an agent or assistant grounded context without dumping the entire page.
CRDS (Cluster Replicated Data Store) (crds) Category: Core Protocol Definition: Cluster Replicated Data Store—the data structure underlying Solana's gossip protocol. CRDS uses a push/pull model where validators exchange versioned, signed records (contact info, vote state, epoch slots). Records have wallclock timestamps and are pruned when stale. Each record is keyed by (source pubkey, label). Aliases: CRDS Related: Gossip Protocol, Cluster
Use glossary context, relationships, mental models, and builder paths to get structured answers instead of generic chat output.
Optional: paste Anchor, Solana, or Rust code so the Copilot can map primitives back to glossary terms.
The Copilot will answer using the current term, related concepts, mental models, and the surrounding glossary graph.
These branches show which concepts this term touches directly and what sits one layer beyond them.
A protocol used by validators to exchange metadata and discover peers without relying on a central server. Gossip propagates contact info, vote status, shred versions, epoch slots, and other protocol messages via CRDS. It does not carry transaction or block data—those use TPU and Turbine respectively.
A set of validators working together to maintain a single Solana ledger. Solana operates three main clusters: mainnet-beta (production), devnet (development), and testnet (stress testing). Each cluster has its own genesis block, token supply, and independent validator set. Validators in a cluster discover each other via the gossip protocol.
These are the next concepts worth opening if you want this term to make more sense inside a real Solana workflow.
These entries are easy to mix up when you are reading quickly, prompting an LLM, or onboarding into a new layer of Solana.
A set of validators working together to maintain a single Solana ledger. Solana operates three main clusters: mainnet-beta (production), devnet (development), and testnet (stress testing). Each cluster has its own genesis block, token supply, and independent validator set. Validators in a cluster discover each other via the gossip protocol.
The high-throughput multicast network used for propagating block data (shreds) and forwarding transactions between validators. The data plane uses Turbine for block propagation and QUIC for transaction submission, optimized for bandwidth and low latency. It operates independently from the control plane (gossip), which handles metadata and coordination.
Glossary entries become useful when they are connected. These links are the shortest path to adjacent ideas.
A protocol used by validators to exchange metadata and discover peers without relying on a central server. Gossip propagates contact info, vote status, shred versions, epoch slots, and other protocol messages via CRDS. It does not carry transaction or block data—those use TPU and Turbine respectively.
A set of validators working together to maintain a single Solana ledger. Solana operates three main clusters: mainnet-beta (production), devnet (development), and testnet (stress testing). Each cluster has its own genesis block, token supply, and independent validator set. Validators in a cluster discover each other via the gossip protocol.
These entries live beside the current term and help the page feel like part of a larger knowledge graph instead of a dead end.
A clock mechanism that cryptographically proves the passage of time between events. PoH uses a sequential SHA-256 hash chain where each output becomes the next input, creating a verifiable ordering of events without requiring consensus. The leader produces ~400,000 hashes per slot (~400ms), and any validator can verify the sequence in parallel, enabling Solana's high throughput by removing the need for validators to agree on time.
Solana's custom BFT consensus algorithm built on top of Proof of History. Tower BFT uses PoH as a clock to reduce communication overhead in traditional PBFT from O(n²) to O(n). Validators vote on forks with exponentially increasing lockout periods—each consecutive vote doubles the lockout, making rollbacks progressively more expensive. A fork is finalized when it reaches supermajority (66.7%+ of stake).
A time window during which a designated leader validator can produce a block. Each slot lasts approximately 400 milliseconds. Slots are numbered sequentially from genesis and grouped into epochs of 432,000 slots (~2-3 days). Not every slot produces a block—a skipped slot means the leader was offline or too slow.
A set of entries produced by a leader during a single slot. A block contains transactions bundled into entries, each with a PoH hash proving ordering. Blocks are broken into shreds for network propagation via Turbine. Maximum block size is limited by compute units (48M CU cap per block) rather than byte size.