Mempool

A mempool is a node's local, unconfirmed transaction pool. It is not a single, global entity but a collection of individual pools maintained by each validating node. Nodes add valid transactions they receive from peers and broadcast them to their own connections, creating a dynamic, peer-to-peer network of pending transactions. The specific contents can vary between nodes due to network propagation delays and individual node policies.

The mempool creates a real-time fee market where transactions compete for block space. Miners or validators typically select transactions with the highest fee-per-byte or gas price to maximize revenue. Users can attach higher fees to incentivize faster inclusion. This results in a priority queue where lower-fee transactions may experience significant delays during periods of high network congestion.

Transactions enter the mempool after passing initial validation checks (signature, nonce). They remain until:

• Inclusion in a Block: The transaction is mined and confirmed.
• Eviction: Nodes may drop transactions after a Time-To-Live (TTL) period or when the pool reaches its memory limit, typically removing the lowest-fee transactions first.
• Replacement: A new transaction with the same nonce but a higher fee (Replace-By-Fee - RBF) can replace a pending one.

The public visibility of pending transactions in the mempool enables Maximal Extractable Value (MEV) opportunities. Network participants (searchers, validators) can analyze the pool to profit by:

• Frontrunning: Submitting their own transaction with a higher fee to execute before a visible pending trade.
• Backrunning: Executing a transaction immediately after a known pending transaction.
• Sandwich Attacks: Placing orders both before and after a large trade. This creates risks for regular users and is a major area of protocol research.

Mempool behavior is not standardized by protocol. Key differences between node implementations (e.g., Geth, Erigon for Ethereum; Bitcoin Core, Knots for Bitcoin) include:

• Size Limits: Maximum memory or transaction count.
• Eviction Policies: Logic for removing stale transactions.
• Fee Estimation Algorithms: Methods for suggesting optimal gas prices.
• Privacy Features: Some clients may implement measures to reduce MEV exposure. This variance means network-wide mempool state is an approximation.

A transaction's journey from creation to finality begins in the mempool. The key stages are:

• Submission: A user broadcasts a signed transaction to a node.
• Propagation: The node validates and gossips it to peers, populating the network's mempools.
• Selection: Miners or validators pick transactions based on fee priority and other rules.
• Execution & Finality: The transaction is included in a block, executed, and removed from all mempools upon sufficient confirmations.

The mempool is the primary arena for blockchain fee markets. Users compete for block space by attaching a transaction fee (often measured in gas price or sat/vB). During congestion, a backlog forms, creating a priority queue where higher-fee transactions are mined first. This dynamic allows for fee estimation tools that suggest optimal fees based on current mempool depth and desired confirmation speed.

Not all mempools are identical. Key implementation differences include:

• Memory vs. Disk: Most are in-memory for speed, but some (e.g., Bitcoin's) can spill to disk.
• Validation Rules: Nodes may enforce different policies (e.g., minimum fee, standardness checks).
• Topology: Some networks use a gossip protocol for propagation, while others may use direct relay networks to reduce latency and prevent censorship.

The public nature of pending transactions introduces attack vectors. Common threats include:

• Frontrunning: Exploiting visible pending trades in DeFi by submitting a higher-fee transaction.
• Time-Bandit Attacks: Attempting to replace low-fee transactions.
• Denial-of-Service (DoS): Spamming the network with low-fee transactions to bloat the mempool. Mitigations include fee bumping (Replace-By-Fee), private transaction pools (like Flashbots' mev-geth), and stricter mempool admission policies.

In Ethereum and compatible chains, the mempool interacts uniquely with the Ethereum Virtual Machine (EVM).

• Gas Estimation: Transactions specify a gas limit and maxPriorityFeePerGas / maxFeePerGas.
• Simulation: Nodes simulate transactions locally before accepting them to check for validity and potential reverts.
• MEV Opportunity: The public order flow in the mempool is the primary source for Maximal Extractable Value (MEV), leading to specialized searchers and builders who optimize block construction.

A malicious actor observes a pending transaction in the mempool and submits their own transaction with a higher gas fee to be executed first. This is common in DeFi arbitrage and NFT minting.

• Sandwich Attacks: A specific form where the attacker places one transaction before and one after the target transaction to profit from price slippage.
• Time-Bandit Attacks: Attackers reorganize blocks to insert their own transactions, exploiting the probabilistic nature of block finality.

Attackers flood the mempool with low-fee, computationally heavy, or data-intensive transactions to clog the network.

• Impact: Increases latency for legitimate users, forces higher gas fees, and can destabilize node operations by consuming memory and CPU.
• Example: Sending many contract creation (CREATE) transactions, which are more expensive for nodes to validate than simple transfers.

An unconfirmed transaction's signature or other data can be altered without invalidating it, creating a different transaction ID (txid).

• Risk: Can be used to double-spend or create confusion in payment tracking systems.
• Mitigation: Modern networks use SegWit (Segregated Witness) or other techniques to fix transaction malleability, making txids immutable once signed.

The public mempool broadcasts transaction details—sender, recipient, amount, and smart contract calls—before confirmation.

• Wallet Fingerprinting: Analysts can link addresses by timing and gas patterns.
• Strategy Sniping: In on-chain games or DeFi, an opponent's pending move can be seen and countered.
• Solution: Services like Flashbots' SUAVE or private transaction relays (e.g., Tornado Cash for ETH) submit transactions directly to miners/validators, bypassing the public mempool.

An attacker isolates a node from the honest network, feeding it a manipulated view of the mempool.

• Mechanism: The attacker connects many malicious peers to the target node, controlling all incoming transaction data.
• Goal: To enable double-spending by convincing the node that an invalid transaction chain is valid, or to censor specific transactions.

Attackers can artificially inflate the perceived network congestion by spamming the mempool, causing wallet fee estimation algorithms to recommend excessively high fees.

• Result: Users overpay for transaction inclusion.
• Defense: Wallets and RPC services use historical data smoothing, dedicated fee estimation APIs, and analyze pending transactions from multiple peers to filter spam.

A transaction's journey from creation to finalization on-chain. Key stages include:

• Creation & Signing: A user creates and cryptographically signs a transaction.
• Broadcast: The transaction is sent to a network node and enters the mempool.
• Propagation: Nodes share the transaction across the peer-to-peer network.
• Mining/Validation: A validator selects it from the mempool, includes it in a block, and executes consensus.
• Confirmation: The block is added to the chain, providing finality.

Mechanisms used to prioritize transactions within a mempool and compensate validators for computation.

• Gas (Ethereum): A unit measuring computational work. Users specify a gas limit and gas price.
• Base Fee (EIP-1559): A network-determined, burned fee per unit of gas.
• Priority Fee (Tip): An additional fee paid to the validator to incentivize inclusion.
• Fee Estimation: Wallets analyze the current mempool to suggest optimal fee rates for timely confirmation.

Exploitative behaviors that leverage mempool visibility.

• Frontrunning: A bot sees a pending transaction (e.g., a large DEX trade) in the public mempool and submits its own transaction with a higher fee to execute first, profiting from the anticipated price impact.
• Maximal Extractable Value (MEV): The total value that can be extracted from block production beyond standard block rewards, often by reordering, inserting, or censoring transactions from the mempool. Strategies include arbitrage, liquidations, and sandwich attacks.

The process by which a newly created block is transmitted across the network. This directly impacts mempool state.

• When a validator publishes a block, nodes receive the list of included transactions.
• Nodes then remove these transactions from their local mempools, preventing double-spending.
• Fast propagation protocols (e.g., Graphene, Compact Blocks) are crucial for reducing orphan/uncle blocks and ensuring mempool synchronization.
• Slow propagation can lead to network forks and mempool inconsistencies.

How applications interact with a blockchain's mempool.

• JSON-RPC: The standard protocol for querying node data. Key methods include eth_sendRawTransaction (to broadcast) and eth_getBlockByNumber (to see included tx).
• Full Nodes: Maintain a complete copy of the blockchain and a full mempool, serving RPC requests.
• Archive Nodes: Full nodes that retain all historical state, useful for analyzing past mempool data.
• Light Clients: Do not store the full chain or mempool; they rely on full nodes for data.

Solutions designed to address mempool limitations like congestion and frontrunning.

• Private Transaction Pools (Flashbots, Taichi): Allow users to submit transactions directly to validators without exposing them to the public mempool, mitigating frontrunning.
• Off-Chain Order Books (DEXs like dYdX): Match orders off-chain and settle batches on-chain, reducing mempool load.
• Layer 2 Rollups: Execute transactions off-chain in batches, submitting only compressed proofs to Layer 1, drastically reducing mainnet mempool contention.