Blockchains require scaling mechanisms to handle increased transaction volumes. Layer 2 solutions offload transaction processing from Layer 1, with three primary architectures: rollups, sidechains, and appchains.

Rollups execute transactions off-chain and batch them for submission to Layer 1, reducing data footprint and fees. Optimistic rollups assume transactions are valid by default, while zero-knowledge rollups use cryptographic proofs—each with different latency and computational tradeoffs.

Sidechains operate as independent blockchains with their own consensus, connected to Layer 1 via two-way bridge. They offer greater speed and flexibility but carry security risk: validators must be trusted separately from Layer 1.

Appchains are purpose-built for specific applications or protocols. They provide dedicated throughput and customizable execution environments—Dydx's move to sovereign appchain, for example, optimized for derivatives. The tradeoff is fragmented liquidity.

Layer 2 designs also vary between monolithic and modular architectures. Monolithic L2s handle execution, data availability, and consensus together. Modular designs separate these functions—Arbitrum handles execution while Ethereum provides data availability and settlement, reducing L2 validator burden.

Most Layer 2s issue native tokens for transaction fees, governance, and validator incentives. Token emissions align developer and user behavior with protocol growth but create sell pressure and governance concentration risk.

Developers and users evaluate L2s on transaction speed, fee structure, and security guarantees. No solution optimizes all three: faster execution typically increases costs or requires weaker security assumptions than the base layer.