The proposal replaces Ethereum’s linear nonce with a key-independent nonce system.
Buterin advises that without this change, nodes would need to download up to 16TB to function.
Vitalik Buterin, along with three other collaborators in the Ethereum ecosystem, developed EIP-8250, a proposal that promises to eliminate a design block that holds back the network’s privacy protocols. According to Buterin, the problem is not only technical, but if not resolved at scale, it could compromise the decentralization of Ethereum.
The blockage is in how Ethereum orders transactions. Each address operates with a linear noncean internal counter that forces transactions to be processed in strict queue, so if one is delayed all subsequent ones from the same sender are paralyzed. This affects privacy protocols because they channel the operations of many users through a single shared address and the delay of one blocks all others.
The EIP-8250, presented this May 5, proposes replace that single counter with keyed noncesthat is, independent lanes through which each transaction can circulate without queuing behind the others.
In privacy protocols, Those lanes would be identified with a structure called nullifiera one-time cryptographic fingerprint that confirms that a withdrawal was not executed before. Thus, multiple users could withdraw funds at the same time from a shared address without blocking each other.
The EIP-8250, which must now be reviewed by the community, points to the Ethereum Hegota update, without a confirmed date, and also depends on EIP-8141another proposal in which Vitalik participated and which is also not implemented.
Vitalik’s underlying argument: decentralization at scale
Buterin, in a post on “a possible first step towards a new state scaling strategy for Ethereum.”
The state of Ethereum is the real-time record of everything that exists on the network: balances, active contracts, variables stored by each application. The more that record grows, It is more difficult for small nodes to continue participating.
Buterin projected that at 2,000 transactions per second of privacy trading over eight years, Ethereum would accumulate 500 billion nullifiers stored on the network. The challenge, he noted, is that these nullifiers “cannot be pruned”, that is, they cannot be removed from the history once registered.
Manage that amount of nullifiers within Ethereum’s general ledger, where any data can be read or modified at any time, would require any node wanting to build blocks to download 16 terabytes of data, which would require would expel participants with modest hardware from the processButerin warned. The co-founder of Ethereum calls these types of records “dynamic states,” and on an extreme scale he considers them incompatible with decentralization.
Finally, faced with this potential problem that threatens small nodes (and therefore decentralization), the Russian-Canadian programmer argued that managing the nullifiers in a dedicated space, separated from the general state, it would allow scaling techniques impossible in the current scheme to be applied.
Among those techniques, two stood out in particular. First, the sharding (fragmentation) of nullifierswhere each node stores only a fraction of the totalso that each one saves only a portion and consults his peers when he needs to verify the rest.
Secondly, Vitalik also highlighted probability filters (bloom filters)mathematical structures that allow verifying whether a nullifier It was already used consuming just 8 bits per entry, instead of consulting the entire register.
The conclusion of the Ethereum co-founder is that “the fully dynamic state is much more difficult to manage at extreme scale than the more controlled and restricted state.” Move most usage toward specialized forms of state, he argued, would allow Ethereum to be kept decentralized and scalablereserving dynamic state for applications that really need it.