Layer 2 Networks Explained
Layer 2s exist because a blockchain that every participant must verify cannot also process thousands of transactions per second. The fix is to do the work elsewhere and prove it.
Key points
- Base layers trade throughput for decentralisation and security — the scalability trilemma.
- Rollups execute transactions off-chain and post compressed data back to the base layer.
- Optimistic rollups assume validity and allow challenges; ZK rollups prove it cryptographically.
- Bridges between layers have been the largest single source of losses in crypto.
Why base layers are slow
A public blockchain's security comes from every node independently verifying every transaction. That is also its throughput ceiling: the network can only go as fast as ordinary participants can keep up with.
Raising the limit — bigger blocks, faster block times — increases the hardware and bandwidth required to run a node. Fewer people run nodes, validation concentrates among a smaller group, and the property that made the system trustworthy erodes. This tension between scalability, security and decentralisation is usually called the scalability trilemma.
The result in practice: Bitcoin handles around 7 transactions per second, Ethereum's base layer roughly 15–30. During periods of heavy demand, fees rise to ration the available space, and Ethereum transaction fees have at times exceeded $50 for a simple transfer.
How rollups work
A rollup executes transactions on a separate network and periodically posts a compressed summary back to the base layer, which remains the ultimate authority on the state.
- Users transact on the layer 2, which has its own sequencer and far higher capacity.
- Many transactions are batched together and compressed.
- The batch, plus a proof or a challenge window, is posted to the base layer.
- The base layer stores the data and provides the security guarantee.
The cost saving comes from amortisation: one base-layer posting covers thousands of layer 2 transactions, so each user pays a fraction of a base-layer fee. Typical savings are 10–100×, and Ethereum's introduction of dedicated data space for rollups in 2024 reduced costs further.
Crucially, the funds remain secured by the base layer. If the layer 2's operators disappear, users can in principle withdraw directly from the base layer using the posted data — though the practicality of this varies considerably between implementations.
Optimistic and zero-knowledge
| Optimistic rollups | ZK rollups | |
|---|---|---|
| Assumption | Batches are valid unless challenged | Validity proved cryptographically |
| Withdrawal to base layer | ~7 days (challenge window) | Minutes to hours |
| Computational cost | Low | High — proof generation is expensive |
| Compatibility | Very close to Ethereum | Historically harder, now improving |
| Maturity | Longer track record | Newer, advancing quickly |
Optimistic rollups post batches and assume they are correct. Anyone can submit a fraud proof during a challenge period, typically seven days, after which the batch is final. The design is simple and compatible, and the cost is that direct withdrawals to the base layer take a week.
Zero-knowledge rollups post a cryptographic proof that the batch was executed correctly. The base layer verifies the proof, so no challenge period is needed and withdrawals are fast. The trade-off is that generating proofs is computationally demanding and building a fully compatible execution environment has been technically hard — though this gap has narrowed substantially.
Risks worth knowing
- Bridges. Moving assets between layers has produced the largest losses in crypto: the Ronin bridge lost around $620 million and Wormhole around $320 million. Use canonical bridges where possible and treat third-party bridges as high risk.
- Centralised sequencers. Most rollups currently run a single sequencer operated by the development team. It can censor or reorder transactions, and if it goes offline the network stops. Decentralising this is work in progress across the sector.
- Upgradeable contracts. Many layer 2 contracts can be upgraded by a multisig, which is a trusted party in a system marketed as trustless. Check who holds those keys.
- Withdrawal delays. Seven days is a long time in a falling market. Fast-exit services exist and charge for it, introducing another counterparty.
- Fragmentation. Liquidity and users split across many layer 2s, which degrades the experience and multiplies the bridging required.
The independent site L2Beat publishes risk assessments covering exactly these dimensions, and is worth consulting before committing significant funds to any particular network.
Frequently asked questions
Are layer 2 networks as secure as the base layer?
Not quite, and the gap varies. They inherit the base layer's security for data availability and settlement, but add their own trust assumptions — sequencer centralisation, upgradeable contracts, and bridge implementations. The honest framing is 'much more secure than a separate chain, somewhat less than the base layer'.
Do I need to bridge assets to use one?
Usually yes, though many exchanges now allow direct withdrawal to major layer 2 networks, which avoids bridging entirely and is generally both cheaper and safer.
What is the difference between a layer 2 and a sidechain?
A layer 2 derives its security from the base layer by posting data and proofs to it. A sidechain has its own independent consensus and validator set, so its security is entirely its own. Sidechains are often faster and cheaper, with correspondingly weaker guarantees.