How the Mempool Works
The mempool (short for "memory pool") is where nodes store unconfirmed transactions. When you broadcast a transaction, it propagates across the network. Each node receives it, validates that it's properly signed and follows consensus rules, then holds it in memory until a miner or validator includes it in a block.
The mempool isn't a single shared location—every node maintains its own local version. This means different nodes can have slightly different mempools at any given moment, depending on network latency and which transactions they've received. But they converge quickly as transactions propagate.
When a miner or validator builds a new block, they select transactions from their mempool. The selection process is economic: they prioritize transactions with higher fees because those maximize their revenue. Once a transaction is included in a block and that block is confirmed, nodes remove the transaction from their mempools. It's no longer "pending"—it's settled.
The mempool has no fixed size limit across all blockchains, but individual nodes can configure their own limits. Bitcoin Core's default mempool size is 300 MB. When a node's mempool fills beyond its limit, it starts dropping the lowest-fee transactions to make room for new ones. Those dropped transactions don't disappear from the network entirely—they might still exist in other nodes' mempools—but if enough nodes drop them, they effectively vanish and need to be rebroadcast.
Where Constraints Live
The primary constraint is block space. Bitcoin produces a new block roughly every 10 minutes with a size limit around 1-4 MB depending on transaction structure. Ethereum targets 12-second block times with a gas limit around 30 million per block. When transaction demand exceeds available block space, the mempool fills.
Fee markets emerge from this constraint. Users signal how much they're willing to pay to have their transaction processed soon. Miners and validators respond by prioritizing the highest-paying transactions. During congestion, fees spike—not because the network is broken, but because users are bidding for scarce blockspace.
There's also a timing constraint: transactions don't wait forever. If a transaction sits in the mempool too long without being included, nodes may eventually drop it. Bitcoin Core, for example, drops transactions after about two weeks by default. The exact timing varies by node configuration.
Network propagation creates another subtle constraint. A transaction needs to reach enough nodes—especially mining or validating nodes—to have a realistic chance of inclusion. If your transaction doesn't propagate widely, it might sit in a few nodes' mempools but never reach the entities building blocks.
What's Changing
Layer 2 solutions have fundamentally shifted where transaction activity happens. Ethereum rollups like Arbitrum and Optimism process transactions off-chain and only post batched data to Ethereum's Layer 1. This means individual user transactions no longer hit Ethereum's mempool directly—only the rollup's settlement transactions do. The result: Layer 1 mempool congestion has declined even as overall transaction volume has grown.
EIP-1559 changed Ethereum's fee mechanism in 2021. Instead of a pure auction, there's now a base fee that adjusts algorithmically based on recent block fullness, plus an optional priority fee (tip). This makes mempool behavior more predictable—you can estimate the base fee needed for inclusion rather than guessing at auction dynamics.
Priority fee markets are evolving. On Solana, recent congestion led to the development of more sophisticated fee markets where users can pay for transaction priority. This is similar to what Ethereum has had with tips, but the implementation details matter for how reliably your transaction gets processed.
Account abstraction proposals (like ERC-4337) could eventually allow transactions to pay fees in tokens other than the network's native currency. This would change mempool economics by enabling fee payment flexibility without requiring users to hold ETH or BTC first.
What Would Confirm Continued Relevance
Sustained mempool activity above minimal levels indicates genuine economic usage, not just speculative trading. If mempools remain consistently full with diverse transaction types (DeFi interactions, NFT transfers, payments, smart contract deployments), that signals real demand for blockspace.
Layer 2 dominance continuing while Layer 1 mempools stabilize would confirm the rollup-centric scaling approach is working as designed. You'd see high Layer 2 transaction volumes with modest Layer 1 mempool pressure as rollups efficiently batch settlements.
Improved mempool infrastructure—like better fee estimation tools, transaction acceleration services, or standardized transaction replacement mechanisms—would signal that the ecosystem is maturing around mempool dynamics rather than trying to eliminate them.
What Would Invalidate Current Mempool Dynamics
Blockchain designs that eliminate mempools entirely would change the picture. Some newer chains experiment with different transaction ordering mechanisms that don't rely on a mempool waiting area. If these prove superior, the mempool model might become obsolete for new networks.
Regulatory restrictions forcing transaction censorship would break the permissionless mempool model. If validators or miners were required to filter transactions based on sender address or transaction content, mempools would shift from neutral waiting areas to curated queues.
Persistent spam attacks that make mempools permanently unusable would force architectural changes. If attackers can cheaply flood mempools with junk transactions that overwhelm fee markets, networks would need to redesign how they handle unconfirmed transaction queuing.
Alternative fee markets that bypass the mempool—like private transaction relays or sealed-bid auctions—becoming dominant would reduce the mempool's importance. MEV (maximal extractable value) markets already operate partly outside public mempools via private order flow.
Timing Perspective
Now, the mempool is the active infrastructure determining transaction confirmation times and fees. If you're using Ethereum Layer 1 or Bitcoin, understanding mempool state tells you what fee to set and when to submit transactions to avoid overpaying during congestion.
Next, watch how Layer 2 adoption affects Layer 1 mempool dynamics. As more activity migrates to rollups, Layer 1 mempools should experience less sustained congestion, making fees more predictable and confirmation times more consistent.
Later, alternative consensus mechanisms and transaction ordering systems may reduce or eliminate traditional mempool architecture. But for Bitcoin and Ethereum—the two largest networks by economic activity—mempools remain core infrastructure for the foreseeable future.
Boundary Statement
This explanation covers the mempool mechanism and its role in transaction processing. It does not constitute advice about optimal fee settings, which depend on current network conditions and individual urgency requirements. Fee estimation tools provide real-time guidance; this post explains why those tools exist.
The mempool works as described. Whether current fee levels represent an acceptable user experience depends on factors outside this scope—primarily whether Layer 1 settlement is necessary for your use case or whether Layer 2 alternatives suffice.
