Block Propagation is the spread of a new block between computers on the Bitcoin network. Delivery provides data; the receiving node still has to check whether the block is valid and belongs in its accepted history.
Imagine Anna runs a mining operation and has just found a block containing transactions. A block groups transactions into a batch added to the shared history. Bitcoin has no central office that sends it to everyone at once. Her node tells its neighbours: the computers it is connected to. The news travels onward, reaching different people at different times. This journey is Block Propagation. Its speed affects how long others keep working from the previous version of the history. [1] [6]
Peter runs a Full Node at home: software that checks Bitcoin blocks and transactions against the rules. A notice about a new block is not enough. He needs its complete contents to verify it. He can download them all, or reconstruct them using the message and transactions he already holds. Think of receiving a list of book pages and finding them in your own library: less travels, but the completed book must be the same. [1] [7]
A block has a short summary called its Block Header. Among other things, it links to the previous block and allows checking Proof of Work, evidence of the computational work mining requires. Miners find this proof through repeated calculations; others can quickly check the result. This is a useful first check, not proof that every payment is valid. The header alone cannot tell Peter whether transactions have valid signatures or really spend available, unspent funds. A digital signature provides proof of authorization for a payment using a private key, a secret numerical value, without revealing that key. A UTXO is an unspent transaction output: a record of an amount and the conditions for spending it. [6] [8]
Peter received many transactions before the block was mined. He keeps them in his Mempool, a local collection of unconfirmed transactions. Compact Blocks, specified in BIP 152, take advantage of this. Instead of sending every transaction again, the message mainly carries short identifiers, alongside some complete transactions. Peter puts the familiar pieces into the correct order. The network message gets smaller, not the block being checked; its validity rules stay the same. [2] [7]
There is no single worldwide Mempool. Each node may know a slightly different set of transactions, so Peter can be missing some. He requests the missing data, or falls back to downloading the full block. That requires another trip across the network and back. It is no reason to guess the contents or skip verification. Actual savings therefore also depend on how many transactions the receiver already knows. [2] [7]
The faster BIP 152 mode allows forwarding before every transaction check is finished. This does not mean blindly sending anything: the sender must check, among other things, that the block builds on the validated chain, meets Proof of Work requirements and has a header that commits to the transmitted transactions. Relay and the remaining checks can then overlap. A Full Node still fully verifies the block before accepting it into its active history. [2] [4] [7]
Before Anna’s block reaches everyone, another miner may find a block extending the same predecessor. Part of the network briefly sees one valid branch and part sees another. Subsequent accumulated work decides between valid branches, not a vote of computers. A valid block left behind is called a Stale Block. Faster propagation narrows the window for this race, but cannot eliminate it. [6] [8]
If Anna’s block spreads slowly, others have more opportunity to extend a competitor instead, leaving her block outside the accepted chain. It then earns her no reward in that history. Good connectivity is therefore more than a convenience. Its advantage can encourage mining to concentrate around better-connected operators. A slow link alone, however, is neither a measurement of centralization nor evidence of an attack. [7] [8]
FIBRE is one example of a specialized system for fast block relay. It aims to reduce network delays, including those caused by lost packets: small pieces of the transmitted message. Such a route can deliver a block sooner, but its operator does not get to decide validity for Peter. His Full Node must check the same rules, whichever route brought the data. [3] [9]
Network operation also has to account for data use, privacy and resilient connections. Relying on a single route is a weakness, however fast it is. Different improvements also do different jobs. BIP 330, a building block of Erlay, helps neighbouring nodes find differences between sets of transaction announcements. It does not replace BIP 152 block relay. The shared aim is more efficient communication while each node retains its own checks. [1] [5]
For the clearest picture, read this entry together with Compact Blocks, Selfish Mining, Eclipse Attack, Initial Block Download, Bitcoin, Full Node. The reverse links also lead from Sybil Attack, Eclipse Attack, Selfish Mining, Compact Blocks.