> ## Documentation Index
> Fetch the complete documentation index at: https://mintlify.com/bitcoin/bitcoin/llms.txt
> Use this file to discover all available pages before exploring further.

# Blockchain Storage and Validation

> Understanding how Bitcoin Core validates and stores the blockchain data structure

## Overview

Bitcoin Core maintains the blockchain as a tree-shaped data structure, with each block containing a cryptographically-secured link to its predecessor. The implementation manages block storage, validation states, and efficient access through sophisticated indexing mechanisms.

## Block Structure

Every block in Bitcoin consists of two main components:

### Block Header

The block header (80 bytes) contains:

* **nVersion** (4 bytes): Block version number indicating protocol rules
* **hashPrevBlock** (32 bytes): SHA256 hash of the previous block header
* **hashMerkleRoot** (32 bytes): Root of the merkle tree of all transactions
* **nTime** (4 bytes): Unix timestamp when the miner started hashing
* **nBits** (4 bytes): Compact representation of the target difficulty
* **nNonce** (4 bytes): Counter used for proof-of-work computation

### Block Body

The block body contains the vector of transactions (`vtx`), where:

* The first transaction must be a coinbase transaction creating new coins
* Subsequent transactions spend previously unspent outputs
* All transactions are included in the merkle tree whose root is in the header

## Validation States

Bitcoin Core tracks multiple validation levels for each block using the `BlockStatus` enumeration:

<CodeGroup>
  ```cpp Validation Levels theme={null}
  BLOCK_VALID_TREE         = 2  // Headers validated
  BLOCK_VALID_TRANSACTIONS = 3  // Transactions structurally valid
  BLOCK_VALID_CHAIN        = 4  // No double-spends, valid outputs
  BLOCK_VALID_SCRIPTS      = 5  // All signatures verified
  ```
</CodeGroup>

### Progressive Validation

1. **VALID\_TREE**: All parent headers found, difficulty matches, timestamp >= median of previous 11 blocks
2. **VALID\_TRANSACTIONS**: First tx is coinbase, no duplicate txids, merkle root matches
3. **VALID\_CHAIN**: Outputs don't overspend inputs, no double-spends, coinbase maturity enforced
4. **VALID\_SCRIPTS**: All script signatures and witness data verified

<Note>
  Blocks can be at different validation levels. A block might have VALID\_TRANSACTIONS but not yet VALID\_SCRIPTS if script verification hasn't completed.
</Note>

## Storage Architecture

### Block Files (blk\*.dat)

Bitcoin Core stores blocks in sequential files:

* Located in the `blocks/` directory
* Named `blkNNNNN.dat` where NNNNN is a 5-digit sequence number
* Each file is approximately 128 MiB
* Blocks are stored in network serialization format
* Files also contain a 4-byte magic identifier and size prefix

### Undo Files (rev\*.dat)

For each block file, there's a corresponding undo file:

* Contains the UTXOs consumed by transactions in the block
* Enables efficient blockchain reorganizations
* Required for rolling back blocks during reorgs
* Stored in custom compressed format

### Block Index (LevelDB)

The `blocks/index/` directory contains a LevelDB database that indexes:

* Block hash to file position mapping
* Block metadata (height, chainwork, validation status)
* Block tree structure (parent/child relationships)
* Transaction counts and timestamps

<Tip>
  The `-blocksdir` option can specify an alternate location for block storage, useful for systems with multiple drives.
</Tip>

## Chain State (UTXO Set)

The `chainstate/` directory contains a LevelDB database representing the **UTXO set** - all currently unspent transaction outputs.

### Coin Structure

Each UTXO ("Coin" in the codebase) contains:

```cpp theme={null}
class Coin {
    CTxOut out;              // The output (amount + scriptPubKey)
    unsigned int fCoinBase;  // Is this from a coinbase?
    uint32_t nHeight;        // Height when included in blockchain
};
```

### Caching Architecture

Bitcoin Core uses a multi-level caching system:

1. **In-memory cache**: Fast access to recently used UTXOs
2. **LevelDB database**: Persistent storage of the complete UTXO set
3. **Cache flushing**: Periodic writes to disk to maintain consistency

## Block Index and Chain Organization

The `CBlockIndex` class represents each block in the block tree:

* **phashBlock**: Pointer to the block's hash
* **pprev**: Pointer to the parent block index
* **nHeight**: Height in the main chain (genesis = 0)
* **nChainWork**: Total work in the chain up to this block
* **nTx**: Number of transactions in the block
* **nChainTx**: Cumulative transactions up to this block

### Active Chain

The "active chain" is the path from the genesis block to the current tip with the most accumulated proof-of-work. During reorgs, the active chain can change if a competing chain achieves more work.

## Pruning

Bitcoin Core supports blockchain pruning to reduce disk usage:

* Retains at least the most recent **288 blocks** (MIN\_BLOCKS\_TO\_KEEP)
* Deletes old block files while maintaining the UTXO set
* Maintains the complete block index for all headers
* Minimum 550 MiB required for block and undo files

<Warning>
  Pruned nodes cannot serve historical blocks to other nodes and cannot rescan the wallet before the pruning point.
</Warning>

## Time Constraints

Bitcoin enforces temporal consensus rules:

* Block timestamp must be greater than the median of the previous 11 blocks
* Block timestamp cannot be more than **2 hours in the future** (MAX\_FUTURE\_BLOCK\_TIME)
* These rules prevent timestamp manipulation attacks

## Reorganization Handling

When a competing chain becomes longer:

1. Identify the fork point (common ancestor)
2. Disconnect blocks from the old chain back to the fork
3. Use undo data to restore spent UTXOs
4. Connect blocks from the new chain
5. Update the UTXO set based on new chain's transactions

## Assumevalid and Checkpoints

### Assumevalid

Bitcoin Core can skip signature verification for blocks before a certain hash:

* Defined in `defaultAssumeValid` consensus parameter
* Only skips script verification, not other validation
* Updated with each release to a recent block
* Can be disabled with `-assumevalid=0`

<Info>
  Assumevalid significantly speeds up initial block download while maintaining security through proof-of-work verification.
</Info>

## Related Concepts

* [Consensus Rules](/concepts/consensus) - The validation rules enforced during block verification
* [Transactions](/concepts/transactions) - The structure and lifecycle of Bitcoin transactions