week four scaling blockchains

What problem are L1 and L2 solutions trying to solve?

They try to make blockchains faster and more scalable.

If a blockchain is slow but already secure, what are two broad ways to scale it?

Upgrade the base chain itself, or build another system on top of it.

What is a Layer 1 solution?

A Layer 1 solution improves or changes the base blockchain itself.

What does “base chain” mean?

The main blockchain that directly handles consensus, execution, and data availability.

What are examples of Layer 1 solutions?

Alternative blockchains, sharding, and Ethereum Improvement Proposals like the Merge.

What is an alternative blockchain?

A separate base blockchain designed with different tradeoffs, often to be faster or cheaper.

Is creating a new faster blockchain an L1 or L2 solution?

It is a Layer 1 solution.

Is upgrading the existing blockchain infrastructure an L1 or L2 solution?

It is a Layer 1 solution.

What was the Ethereum Merge?

The Merge, or EIP-3675, shifted Ethereum from Proof of Work to Proof of Stake.

Why is the Merge considered a Layer 1 change?

Because it changed Ethereum’s base consensus mechanism.

What is a Layer 2 solution?

A Layer 2 solution is built on top of a base chain to handle transactions faster while still relying on the base chain for security.

What does an L2 rely on the L1 for?

Security and final settlement.

What are examples of Layer 2 solutions?

Rollups, state channels, and plasma.

Are rollups L1 or L2 solutions?

Rollups are Layer 2 solutions.

Are state channels L1 or L2 solutions?

State channels are Layer 2 solutions.

Are plasma systems L1 or L2 solutions?

Plasma systems are Layer 2 solutions.

Simple difference between L1 and L2?

L1 changes the base chain; L2 builds on top of the base chain.

What are the three main layers of a blockchain?

Consensus, execution, and data availability.

What is the consensus layer responsible for?

Agreeing on which transactions or blocks are part of the chain and in what order.

What is the execution layer responsible for?

Running transactions and updating the blockchain state.

What is the data availability layer responsible for?

Making sure the data needed to verify blocks is available.

Why can’t you scale only one blockchain layer and expect the whole chain to be fast?

The whole blockchain is limited by its slowest layer.

What does it mean that the blockchain is “bottlenecked” by the slowest layer?

Even if some parts are fast, the slowest part limits the overall speed.

What must happen for the whole chain to become faster?

Consensus, execution, and data availability all need to scale.

What is vertical scaling?

Vertical scaling means making the base chain itself handle more execution.

What are examples of vertical scaling?

Upgrading validators, using faster consensus, and improving state storage.

What does upgrading validators mean in vertical scaling?

It means requiring or enabling validators to use stronger hardware or better infrastructure.

What is the goal of vertical scaling?

To make each part of the base chain more powerful or efficient.

What is the limitation of vertical scaling?

Eventually, there are hardware, cost, and decentralization limits.

What is horizontal scaling?

Horizontal scaling means adding more resources or splitting work across multiple chains or parts of the system.

What are examples of horizontal scaling?

Sharding the blockchain or bridging to other chains.

What does sharding do?

It splits blockchain work across multiple sections, so the whole network can process more activity.

What does bridging to other chains allow?

It lets activity move across multiple blockchains instead of relying on one chain only.

Why is horizontal scaling considered a better long-term goal?

Because vertical scaling eventually hits limits, while horizontal scaling can keep adding capacity.

Simple summary: what is L1 scaling?

Making the base blockchain faster or more capable.

Simple summary: what is L2 scaling?

Building systems on top of the base chain to process transactions faster.

Simple summary: vertical vs horizontal scaling?

Vertical scaling makes one chain stronger; horizontal scaling spreads work across more chains or components.

Simple summary: why is scaling hard?

Because consensus, execution, and data availability all need to keep up.

What are rollups?

Rollups are Layer 2 scaling solutions that execute transactions off-chain and publish compressed transaction data on-chain.

What blockchain layer do rollups mainly improve?

The execution layer.

What does a rollup rely on the base chain for?

Consensus and data availability.

Why are rollups considered Layer 2 solutions?

They run on top of the base chain while still using the base chain for security and data availability.

Why is it important for rollup data to be posted on-chain?

It lets ordinary users verify rollup execution and ensures users can exit back to the main chain.

What does “off-chain execution” mean in rollups?

Transactions are processed outside the main chain instead of being fully executed on L1.

What gets posted on-chain by a rollup?

A compressed version of transaction data and state root updates.

Where is blockchain state commonly stored conceptually?

In a Merkle tree.

Do rollups maintain their own state?

Yes. Rollups maintain their own state separate from the main chain.

What does a rollup publish to a smart contract on the main chain?

The Merkle root of its rollup state.

Is the full rollup Merkle tree stored on-chain?

No. The full Merkle tree is not stored on-chain.

If the Merkle tree is not stored on-chain, how can users verify it?

They can recompute it from the on-chain data.

Why are rollup transactions compressed into batches?

Batching lets many more transactions fit into the same block space.

About how large is a standard ETH transaction?

Around 110 bytes.

About how large can a rollup transaction be?

Around 12 bytes.

Why does smaller transaction size help scaling?

More transactions can fit into the same block size.

What does the smart contract check when it receives a rollup batch?

It checks that the batch’s old state root matches the old state root it already stored.

What happens after the old state root is verified?

The smart contract updates to the new state root claimed in the batch.

Does the L1 smart contract compute the new state root from all transactions?

No.

Why doesn’t the L1 smart contract compute the new rollup state root?

That would defeat the purpose of a rollup, because it would force L1 to redo the expensive execution.

What does the rollup smart contract mainly track?

The rollup’s state roots and batch data.

What is a sequencer in a rollup?

A sequencer is the party that orders transactions and submits batches.

What blockchain problem is sequencer selection similar to?

Selecting a block proposer in proof of stake.

What is the ideal decentralized way to choose batch submitters?

Require submitters to stake the protocol token and slash them if they submit incorrect batches.

What does slashing mean?

Losing part or all of a staked amount for malicious or incorrect behavior.

What do most rollups currently use for sequencing?

Centralized sequencers.

Why are centralized sequencers a tradeoff?

They are efficient, but they create more centralization risk.

What are the two main ways to verify rollup batches?

Fraud proofs and zero-knowledge proofs.

Which rollups use fraud proofs?

Optimistic rollups.

Which rollups use zero-knowledge proofs?

ZK rollups.

What is a fraud proof?

A proof that someone else’s computation was faulty.

What do fraud proof provers need to include?

The contested batch, the correct state root, and information from the state tree.

Why does a fraud proof include state tree information?

So the computation can be recreated and checked.

How do optimistic rollups treat submitted batches at first?

They assume the submitted state root is correct unless challenged.

Who publishes batches in optimistic rollups?

Validators or sequencers publish batches to the chain.

What is the challenge period in optimistic rollups?

A time window where other validators can challenge a batch by submitting a fraud proof.

How long is the optimistic rollup challenge period usually?

Around one week.

What happens if someone posts a valid challenge?

The disputed computation is checked on L1, and the correct state root is determined.

What may happen to the party that loses a fraud-proof challenge?

They may get slashed.

Why can optimistic rollup withdrawals take a while?

Users may need to wait through the challenge period before the system treats the batch as final.

Why are they called “optimistic” rollups?

They optimistically assume batches are correct unless someone proves fraud.

What is the downside of rerunning an entire batch on L1?

It is expensive, so rollups use optimizations to avoid rerunning the whole batch when possible.

What do zero-knowledge proofs allow you to do at a high level?

They allow you to prove that something exists or is true without revealing the thing itself.

What is the basic idea behind a ZK proof?

You can prove knowledge or correctness without exposing the underlying data.

How do ZK rollups use zero-knowledge proofs?

Each batch includes a ZK proof showing that valid signed transactions produce the new claimed state root.

What does a ZK rollup proof prove?

It proves that there exists a set of signed transactions that correctly changes the old state root to the new state root.

Why are ZK proofs useful for rollups?

They can be verified on-chain with minimal computation.

Does Ethereum need to rerun all ZK rollup transactions to verify the batch?

No. It only needs to verify the ZK proof.

How does proof size grow as more computation is bundled into a ZK proof?

The proof gets longer, but its size grows very slowly compared to the amount of computation.

What are examples of popular ZK rollup implementations?

Scroll, Linea, ZKSync, Starkware, and Taiko.

What is the data problem in blockchain scalability?

Blockchains keep storing more data, making it harder for nodes to validate and fetch all the needed information.

Why is growing blockchain data a scalability issue?

If data grows too large, only powerful nodes may be able to store, validate, and fetch it.

How large was Ethereum’s chain data around October 2025 according to the notes?

Around 1.4 TB in chain data alone.

Why does the data problem show that every blockchain layer must scale?

More transactions do not help if nodes spend too much time searching and validating the increased state.

What is data availability?

Data availability is the problem of verifying that transaction data for a new block is public and visible.

Why is data availability important?

Nodes and users need access to transaction data to verify that blocks are valid.

What is Celestia?

Celestia is a blockchain focused on data availability.

What does Celestia mainly handle?

Celestia mainly handles data availability.

What is Celestia like in simple terms?

It is like a public bulletin board where rollups can post transaction data cheaply and securely.

Why is Celestia highly optimized for data retrieval?

Because its main purpose is storing and making data available, not executing transactions.

Does Celestia have an execution layer?

No. Celestia does not have an execution layer.

What does Celestia do with raw data?

It takes raw data without interpreting it and pushes the blob into a block.