The kohaku initiative is expected to add pluggable network-layer privacy support for anonymized transaction routing.
Ethereum’s block building pipeline is at the center of a sweeping set of proposed changes tied to the upcoming Glamsterdam upgrade.
Vitalik Buterin laid out these proposals in a detailed post, covering censorship resistance, MEV mitigation, encrypted mempools, and network-layer privacy.
The roadmap spans both in-protocol and extra-protocol components. Together, these measures aim to reduce builder centralization while protecting users from exploitative transaction ordering across the network.
FOCIL marks one of the first concrete steps toward multi-participant block building. The mechanism requires 16 randomly selected attesters to each nominate a set of transactions.
Those transactions must appear somewhere in the block or the block gets rejected outright. This design ensures censorship resistance even under total builder capture.
Buterin noted that FOCIL prevents exclusion even if a single hostile actor controls all block building. A more expansive version, known as Big FOCIL, has also entered the conversation as a possible next step.
Under that model, FOCILers would collectively cover all transactions in a given block. Each participant handles a subset tied to the sender address’s first hex character by default.
The system avoids excessive duplication through that address-based assignment structure. Only censored transactions risk appearing twice, and then only at the cost of a one-slot delay.
As a result, the builder’s role could narrow considerably over time. It may eventually cover only MEV-relevant activity like DEX arbitrage and state transition computation.
This shift would mark a meaningful reduction in the power held by block builders. The change does not eliminate builders but restructures how much influence they hold.
Buterin framed it as a logical extension of the FOCIL approach. It moves Ethereum further toward a distributed, trust-minimized building process.
Encrypted mempools have emerged as one of the most discussed solutions to toxic MEV. Practices like sandwiching and frontrunning cause direct financial harm to users.
Encryption prevents any party from seeing a transaction’s contents before it lands in a block. That removes the window these attacks currently rely on.
The mechanism works by keeping the transaction hidden until inclusion is confirmed. No one can wrap or reorder it in a hostile way during that window.
However, the technical challenge is non-trivial and remains under active research. Ensuring transaction validity in a mempool-friendly format while preserving encryption is complex.
A further challenge involves guaranteeing that decryption happens only after block finalization. Researchers are exploring several cryptographic techniques to address this.
No single approach has been adopted yet, but progress is ongoing. The goal is a system that is both user-friendly and resistant to manipulation.
The transaction ingress layer has long been an overlooked area in MEV and privacy discussions. A hostile observer that sees a transaction in transit can act before it reaches the chain.
They can sandwich DeFi trades, grief users with invalidating prepended actions, or expose private on-chain activity. The damage can occur well before any on-chain privacy protocol takes effect.
Buterin pointed to growing research in this area. Tools like Tor routing, Ethereum-focused mixnets, and latency-minimized designs like Flashnet are all under consideration.
These approaches aim to anonymize the path a transaction takes from user to block. The kohaku initiative is expected to integrate pluggable support for such protocols.
There is also interest in using the network layer for beneficial pre-inclusion transaction handling. Passive order-matching for DeFi activity could improve execution quality for users.
However, enabling that without opening the door to sandwiching requires careful cryptographic design. That remains an open challenge for researchers working in the space.
Looking further ahead, Buterin outlined a vision of distributed block building modeled loosely on how BitTorrent operates. The idea involves creating new, cheaper transaction types that do not require global state access.
That would allow most activity to flow through a fully distributed pipeline. The current transaction types would remain available but carry a relatively higher cost.
