Skip to main content
Other Assets

Layer 2s and scaling

Pomegra Learn

Layer 2s and scaling

The fundamental constraint facing Bitcoin, Ethereum, and most blockchains is throughput. Bitcoin processes roughly seven transactions per second; Ethereum, roughly 13–15. Visa handles 24,000 transactions per second at peak, and the financial system settles far more than that daily. For cryptocurrencies to function as general-purpose payment or settlement networks, they must solve the scalability trilemma: how to achieve security, decentralization, and high throughput simultaneously.

The answer, pursued aggressively across the industry, is to move computation and settlement off the mainchain while keeping the security guarantees of the mainchain intact. These off-chain or "Layer 2" systems have become the primary mechanism through which Ethereum and other networks achieve application-grade throughput. Understanding how they work—and their tradeoffs—is essential for anyone evaluating a blockchain platform or using one in production.

Layer 2 solutions fall into several families: payment channels (typified by Bitcoin's Lightning Network), sidechains (independent chains that periodically settle to the mainchain), plasma (a design for scalable exits), and rollups (the dominant category today). Rollups further divide into optimistic rollups, which assume batches are valid unless proven otherwise, and zero-knowledge (ZK) rollups, which include cryptographic proofs of correctness in every batch. Each family makes different security and usability tradeoffs. Optimistic rollups like Arbitrum and Optimism prioritize developer simplicity—they run the Ethereum Virtual Machine unchanged—but require a fraud-proof window (typically one to seven days) before deposits are irreversible. ZK rollups like zkSync and StarkNet validate every batch with a proof, enabling faster finality, but require significantly more engineering complexity and currently support fewer applications out of the box.

Bridges—the infrastructure that transfers assets between Layer 2 networks and the mainchain—are critical to usability but introduce their own security risks. A bridge is ultimately a piece of code that locks assets on one chain and mints representations on another; if that code is compromised or its validators are malicious, user funds can vanish. Several major breaches have occurred on bridges, motivating ongoing research into verifiable, light-client-based designs.

Fees on Layer 2s drop dramatically relative to the mainchain—often by 10–100x, depending on the network and current congestion. This cost reduction unlocks use cases that are economically infeasible on mainnet: frequent updates, granular transactions, and applications that simply would not work at $5–50 per transaction. Yet Layer 2s introduce new complexity in applications: users must bridge assets, learn new interfaces, and accept shorter settlement windows (in the case of optimistic systems) or different security assumptions (in the case of sidechains). The industry's focus is now on improving Layer 2 UX, enabling seamless cross-Layer 2 movement, and ensuring that security properties are not lost in the push for speed.

The scalability problem and design space​

Why can't blockchains just increase block size or reduce block time? What constraints limit throughput on the mainchain, and how do Layer 2 architectures relax them?

Production Layer 2 networks​

Which Layer 2 networks have moved beyond testnet, achieved meaningful adoption, and enabled real economic activity? What do Arbitrum, Optimism, zkSync, and Polygon offer developers and users?

Articles in this chapter​