Mainnet Fees Reduced by 47%, A Quick Look at Optimism’s Upcoming Bedrock Upgrade

The “Bedrock” upgrade is expected to bring a 47% fee reduction to the Optimism mainnet, which is a significant leap forward. This not only greatly improves the performance of Optimism, but is also expected to attract more developers and users to join the Optimism ecosystem. The nearly half fee reduction is obviously a big plus, giving Optimism a significant advantage in the highly competitive blockchain market, while also bringing unprecedented convenience and affordability to users. In addition, this upgrade is seen as one of the catalysts for the continued strength of the Optimism ecological market, which indicates that Optimism has a broader and more promising development prospect.

”TL;DR:

• “Bedrock” is a major upgrade of the Optimism mainnet with the goal of significantly reducing usage fees. This change stems from a simple assumption: lower fees equal better user experience.
• Through a series of optimizations and innovations, the Bedrock upgrade is expected to bring about a 47% reduction in protocol costs and security fees for the Optimism mainnet.
• Major optimization measures include:
• Efficiently transmit transaction data: optimize between Layer 1 (one-tier network) and Layer 2 (two-tier network), improve data compression efficiency, and reorganize the data sent from Optimism to L1 to maximize the available data on L1 use of space.
• Improved data compression method: Among various compression algorithms, zlib was finally chosen because it provided good performance for Optimism’s needs.
• Adoption of a new batch processing system: The batch processing system is a line format that converts traditional blocks into batch processing that eliminates additional information as much as possible and retains only a small amount of necessary metadata.
• In addition, the Bedrock version also reduces the Gas fee on Ethereum as much as possible. This is achieved by removing all execution gas and reducing L1 data fees to a theoretical minimum.
• Ultimately, these improvements will help Optimism mainnet users save a lot of transaction costs.
• Once the upgrade is complete, the Optimism team will post an update via the OP Labs twitter account, including post-Bedrock data (which will show how accurate the predictions are), so keep an eye out.

Bedrock Upgrade: A new chapter in Optimism, expected to cost 47% less

Reducing usage costs is a design principle that occupies a large proportion of the Bedrock version. This resulted in a complete re-architecture of how transactions are published to L1 (Layer 1) and vice versa.

As powerful as this approach is, it stems from a very simple assumption: lower fees = good. We want our ecosystem to be the most accessible: the easiest, the most fun, with the lowest construction and transaction costs.

Ultimately, the Bedrock upgrade will bring a 47% reduction in protocol costs and security fees to the Optimism mainnet. Read on to learn how we made it happen.

Where does the fee come from?

Fees for sending transactions on the Optimism mainnet come from two sources: L2 (Layer 2 Network) Execution Fees and L1 Data/Security Fees.

L2 execution fees are similar to how regular transaction fees work on Ethereum, but with the added advantage that execution gas prices on Optimism mainnet are very low since the network is not as congested as L1.

L1 data fee is because all transactions on Optimism are also posted to Ethereum. This step is critical to the security of Optimism, as it means that all data needed to sync Optimism nodes is always publicly available on Ethereum. This makes Optimism an L2. Users on Optimism have to pay a fee to submit their transactions to Ethereum. Since the gas fee on Ethereum is very expensive, L1 data fee occupies a large part of the total transaction cost on the Optimism mainnet.

Looking for improvements

Early in Bedrock’s development, we were able to identify several areas that were causing unnecessary costs to users. In particular, Optimism’s legacy system was designed in such a way that it did not use the L1 data space as efficiently as possible. For example, the pre-Bedrock system that published data to Ethereum simply added as many transactions as possible into a compressed data bundle that could fill a single L1 transaction.

To solve this problem, the most logical place to start is to try improve the efficiency of compressing large amounts of data published to Ethereum. We also found that reorganizing the way data is sent from Optimism to L1 maximizes the use of available data space on L1.

Iteration of repair methods

Once we had a proof-of-concept of submitting transaction groups composed of compressed data that brought us closer to our goal, we focused on optimizing this approach.

Among the key constraints guiding development is the need to split the body of an L2 block into multiple L1 transactions. If someone submits a large transaction to L2, it may need to be split to fit in L1 Calldata. In addition, the algorithm we use to compress the data, the more data input, the better the compression ratio. To address this constraint and maximize the compression ratio, we design a system that can exploit these properties.

Batch and compress data

The key building block of this system is batch processing. Batch is a wire format designed to minimize the cost and software complexity of writing to L1. Traditional blocks are turned into batches, eliminating as much extra information as possible and keeping only the necessary, small amount of metadata.

Next, the list of L2 transactions known as sequence batches is compressed into so-called channels. Each channel has a maximum size (initially about 9.5Mb). These channels will be compressed using a compression algorithm before committing to L1. Compressing large batches into each pass is where we get good compression ratios efficiently.

Channels are further divided into channel frames. This is the part that helps us address the key constraints mentioned above. Slicing channels into frames allows us to handle very large L2 transactions by sending them to L1 in parallel, thus filling up as much as possible with corresponding L1 transactions.

Compression algorithm

When choosing a compression algorithm, we considered candidates such as zstd, brotli, and LZW, but ultimately chose zlib because it provided good performance for our needs.

Good performance in this case means the right balance between a good compression ratio and a good compression and decompression speed that we are trying to achieve. This reflects the typical trade-off in choosing a compression algorithm: speed/compression ratio performance.

Minimize the use of Ethereum Gas

The last part that helps to reduce the cost is that Bedrock removes all execution gas, reducing the L1 data cost to the theoretical minimum. We discuss this in depth in the Bedrock explainer.

Here is an excerpt:

Bedrock removed all execution gas used by the L1 system when sending transactions known as batch transactions. All verification logic that previously occurred on L1’s smart contracts has been moved into block derivation logic (block derivation logic). Instead, batch transactions are sent to a single EOA (Externally Owned Account) on Ethereum known as a batch inbox address.

Batches are still subject to validity checks (i.e. they must be correctly encoded), as are individual transactions within the batch (e.g. the signature must be valid). Invalid batches and invalid individual transactions in otherwise valid batches are considered discarded and are not relevant to the system.

Cost reduction by numbers

After the Bedrock upgrade, we expect a ** protocol cost/security fee reduction of 47%**, including 99% of state chain commitments and 20% of batch commit costs.

We’ll post an update via the OP Labs twitter account after the upgrade, including post-Bedrock numbers (which will show how accurate our predictions are), so keep an eye out.

Next Steps: Fee Optimization and EIP-4844

We take great pride in the designs and solutions we build to keep Bedrock posting costs down. This is a result of our team leaning towards solid engineering fundamentals and executing them cleanly.

Our team is continuing to optimize fees to reduce L2 data fees, so expect to see this in a release after Bedrock. One of the specific things we can do is time batch submissions so that batches are guaranteed to be submitted within a certain time window (say 10 minutes), and we can submit batches when the fee is lowest during those 10 minutes.

Even more exciting is that EIP-4844 will soon be incorporated into Ethereum. When that time comes, the cost of publishing data to L1 will be further reduced.

If this post got you excited about building on Optimism, check out our documentation on how to contribute to the community. Additionally, our Bedrock explainer details many of the exciting features coming with the Bedrock release.