Author: Coindesk

Translation: Shenchao TechFlow

Shenchao Introduction: Zcash can currently process only about 1 private transaction per second, and the goal of the new node Zakura is to achieve 50,000 TPS while maintaining complete privacy—a processing scale comparable to Visa and Mastercard. This is an important signal for everyone concerned about payment privacy and blockchain scalability: privacy and performance are no longer an either-or scenario.

The people rebuilding Zcash have a dream: to reach the level of global payment giants Visa and Mastercard, processing tens of thousands of payments per second while maintaining complete verifiability and strong privacy guarantees.

The first piece of this plan is Zakura, a new full-node software version 1.0.0 released on Wednesday. It is maintained by Sean Bowe, a founding member of Zcash zero-knowledge cryptography, and Dev Ojha, co-founder of Osmosis and current head of Valar Group. Both teams are funded by private ZEC donations rather than by corporations or foundations.

"Our dream is to support global payments. Mastercard and Visa process over 50,000 transactions per second; that is our baseline. Under Zcash's existing cryptographic scheme, this transaction volume requires nodes to process over 500 MB of throughput per second," a blog post stated. "The current technology stack cannot achieve this goal. The cryptographic technology our team is developing can narrow most of the gap."

A full node is a program that stores a complete copy of the blockchain, in this case, the Zcash ledger, and independently verifies whether each transaction complies with the network's rules. Zakura is a forked version of Zcash Foundation's node software Zebra—which means it is rebuilt from the foundation's official code.

The consensus rules are a shared manual of rules executed by each node, deciding which blocks and transactions the entire network accepts as valid. If a node applies different rules, it will fork and stop following the same chain as other nodes.

Pruning, Snapshots, and Compatibility

Zakura can also prune, meaning it deletes old blockchain data that nodes no longer need, significantly reducing disk usage. This compresses the chain to small enough size that the team has released a ready-made copy, about 11 GB after stripping old data, which new nodes can download directly instead of pulling complete history block by block from other nodes.

This allows nodes to go from zero to running in less than two minutes, and the team claims "it is 680 times faster."

Compatibility mode further replicates the interface of the original client zcashd, which will cease its lifecycle on July 18, ensuring that wallets and exchange integrations built for it can continue to function as usual.

Throughput Goals and the Role of Tachyon

The reason for building all this is an arithmetic issue.

Mastercard and Visa process over 50,000 transactions per second, and the team claims this number is "the baseline, not a goal." Zcash's current cryptographic scheme requires nodes to receive and verify over 500 MB of data per second to keep up because each private transaction carries a proof, and the proof is large.

This is roughly equivalent to continuously receiving the data of a full DVD every 10 seconds, and currently, no Zcash software can come close to this level. But the key is to understand the reason for each bottleneck.

Bowe's Tachyon project is addressing this issue through recursive proofs, where one proof can certify the validity of thousands of other proofs, greatly reducing the amount of data that must be checked at the consensus layer.

Under the Tachyon scheme, nodes validate a single proof instead of thousands, and the team claims this will reduce the consensus data requirement from 100 MB per second to 500 MB, which they believe is technically achievable with careful engineering.

Wallet Bottleneck and Valar’s PIR Solution

Wallets face different issues. Because Zcash hides the receiver of transactions, a wallet cannot inquire which transactions belong to it from the server without revealing itself. It must pull everything and test each one, which is why wallet software can achieve no more than about 1 transaction per second.

To eliminate this bottleneck, Valar Group is researching private information retrieval technology, allowing the wallet to retrieve its own data from the server without the server knowing which entries were requested.

Fast Block Propagation

Fast block propagation means broadcasting newly mined blocks as quickly as possible across the blockchain network. Zakura is the software layer responsible for this task.

It must transmit new blocks between nodes fast enough to enable high volumes of proofs and wallet traffic. It is equipped with an experimental system designed to transmit each block to every node within half a second, currently turned off by default.

Upcoming tests of these ideas will arrive in late July. Ironwood, formally named NU6.3, will activate at mainnet block 3,428,143, approximately at 8 AM Eastern Time on July 28, with Zakura supporting it from the day of release.

Bowe stated on July 10 that all major organizations have committed to activate at this height, a week later than originally planned because exchanges and wallet providers requested more preparation time.

The Origin of Ironwood

Ironwood exists because of a vulnerability that nearly destroyed Zcash in June. The so-called shielded pool is the privacy end of the network, with amounts and participants hidden, and zero-knowledge proofs serving as evidence of mathematical computation.

On May 29, Shielded Labs researcher Taylor Hornby discovered that the proof circuit for Orchard (the latest shielded pool) contained a soundness vulnerability that allowed attackers to mint fake ZEC on-chain without a trace. This vulnerability had existed since Orchard's activation in May 2022.

Developers disabled Orchard through an emergency response completed on June 2, then restored it on June 3 with a corrected circuit via the NU6.2 hard fork at block 3,364,600.

The patch could not fix the vulnerability that had been open for four years. Zero-knowledge proofs do not reveal any information beyond verifying this fact, therefore there is no record on-chain of any Orchard transaction transferring content, and no one can prove that fake ZEC was never created.

Ironwood was built to solve this issue. The so-called "rotating door" at the pool boundaries restricts what can leave and enter the pool, leveraging the fact that the amount of ZEC entering and exiting the shielded pool is public, even though transactions within the pool are private. Closing Orchard to stop new deposits, the rotating door becomes the only exit, trapping any internal fake coins.

In simple terms, honest balances can migrate out over time, while fake coins can be prevented from fully exiting and entering the circulating supply. This setup traps any attempts at oversupply within the boundaries, restoring the reliability of the token supply.