Ethereum News
Scaling Ethereum rollups via PeerDAS and BPO

PeerDAS (EIP-7594) changes how Ethereum manages data. Instead of requiring every validator to download every blob, PeerDAS allows sampling. By allowing validators to verify only partial data sets through PeerDAS, the network reduces the bandwidth and storage load on nodes, which enables layer 2 rollups to post more data at lower costs. The network uses Reed-Solomon encoding to expand data size twofold. This allows a node to recover the original data from any 50% of the pieces. The network divides into 128 subnets. A validator staking 32 ETH participates in 8 subnets. Each piece of data is only 2KB. These validators use 16KB of bandwidth per blob instead of the 128KB used previously. The probability of missing data slipping past the sampling mechanism stays between 1 in 10^20 and 1 in 10^24. I consider the efficiency of PeerDAS the best part of the upgrade. The block gas limit also rises from 45 million to 60 million via EIP-7935. Proposers must still have all the blobs in the block to propagate them. Because the proposer does not get the benefits of sampling, the network uses an API called engine_getBlob to let nodes fetch blobs from the execution client mempool.
Rollups like Base and Arbitrum benefit from increased data availability. The network manages this expansion through Blob Parameter-Only (BPO) forks. EIP-7892 allows developers to adjust blob counts without a full hard fork. BPO1 on December 9, 2026, raises the target from 6 to 10 blobs. BPO2 on January 7, 2027, increases the target to 14 and the maximum to 21. This increase moves data capacity from 1.125 MB per block to 2.625 MB per block. I find the staged rollout of these parameters to be a smart way to manage network stability. EIP-7918 ties the blob base fee to execution costs. This prevents the fee market from collapsing to 1 wei. These changes could cut L2 transaction fees by 40% to 60%. Ethereum competes with chains like Solana and Sui that focus on speed. Fusaka changes the narrative by adding low cost to Ethereum security.
| Feature | Specification/Value |
|---|---|
| Default Block Gas Limit (EIP-7935) | 60 million |
| Max Transaction Gas (EIP-7825) | ~16.7 million |
| Max RLP Block Size (EIP-7934) | 10 MB |
| PeerDAS Data Storage per Node | 1/8 of total blob data |
| BPO1 Target Blobs (Dec 2026) | 10 |
| BPO2 Target Blobs (Jan 2027) | 14 |
Scaling capacity for rollups
Rollups process transactions offchain and submit results to the Ethereum mainnet. They must provide the original transaction data to prove they processed transactions correctly. Before the Dencun upgrade, rollups used calldata, which was expensive. Ethereum introduced blobs via EIP-4844 to provide a space for data storage. The current blob target is 6 blobs and the maximum is 9. As demand increases, BPO allows the target to move to 10 and then 14. This expansion moves capacity without requiring a full hard fork. The implementation of these changes allows rollups to carry more data. This helps Ethereum maintain its position in the Web3 stack. I expect the capacity to reach much higher levels as the network adopts these methods.
Validator requirements and user experience
Validators and node operators face new operational realities. PeerDAS reduces disk and download needs by 50% for solo stakers. However, upload bandwidth increases two to three times during activation. I find the increased upload bandwidth requirement for solo stakers to be a blunt drawback that complicates network participation during activation. Large operators with 800 ETH are responsible for about 25 custody columns. They experience 30% higher download bandwidth than pre-Fusaka levels. These operators should maintain at least 100 Mbps upload bandwidth to handle blob propagation. Supernodes with 4096 ETH must store 600 GB of additional blob data. They maintain a sustained download rate near 20 Mbps. You already know the basics of node requirements, so focus on how PeerDAS affects the specific bandwidth ratios.
EIP-7951 adds native support for the secp256r1 curve. This allows wallets to use biometrics like Face ID or Touch ID. EIP-7917 gives proposers more visibility into upcoming slots. This helps reduce MEV variance. EIP-7934 sets a 10 MB limit on RLP encoding to prevent massive blocks. EIP-7825 caps individual transaction gas to 16.7 million. EIP-7939 adds a new CLZ opcode for better smart contract logic. EIP-7823 sets upper bounds for MODEXP precompile inputs to improve security. EIP-7883 increases the gas cost for MODEXP operations.
Future roadmap and network stability
The upgrade builds on the Pectra hard fork. It prepares the network for Glamsterdam in 2026. The development team uses testnets like Hoodi and Sepolia to trial these changes. Hoodi launched in March 2025 as a clean-slate environment for validator testing. Sepolia remains the staging ground for dApp developers. Ephemery provides 28-day validator cycles for specific testing scenarios. The probability of missing data slipping past the sampling mechanism stays between 1 in 10^20 and 1 in 10^24. Will the increased block size eventually force smaller nodes out of the network entirely?