Block.one previewed EOSIO Dawn 4.0’s resource market and scaling design

Block.one published its EOSIO Dawn 4.0 design preview on May 4, 2018, detailing major changes to resource allocation, finality and inter-blockchain communication ahead of the planned EOSIO 1.0 software milestone in June 2018.

The announcement was consequential because it moved several core design choices out of general white-paper language and into a proposed operating model for developers, block producers and exchanges. It was not, however, a production launch or even the formal Dawn 4.0 release. Block.one chief technology officer Daniel Larimer stated that development remained underway in the project’s “slim” GitHub branch and that the team intended to tag the release on May 11, 2018.

That distinction matters. On May 4, EOSIO was still pre-release software, and the performance, security and economic behavior described by Block.one had not been demonstrated on an independently operated public network.

RAM became an internal market

The most concrete change involved RAM, the memory used to hold blockchain state. Block.one said testing had shown that Dawn 3.0’s proportional allocation of RAM to staked token holders could eventually produce shortages. Dawn 4.0 therefore proposed a market using a Bancor-style pricing algorithm, with the system contract buying and selling RAM at prevailing prices.

Under the design, RAM would become more expensive as available capacity declined. Block.one also proposed making RAM non-transferable and charging a 1% fee on purchases and sales. The company acknowledged that market pricing could attract speculation, although it argued that a price signal would encourage holders to release scarce capacity.

The announcement illustrated the model using several assumptions rather than observed market data. It assumed a token price of $20, one terabyte of available RAM and roughly four kilobytes for a new account, producing an estimated initial account-memory cost of about $0.10. Those figures were Block.one’s design calculations, not prices measured on a live EOSIO network.

This approach made infrastructure economics part of application design. Developers would need to minimize persistent state, free unused records and account for a resource price capable of changing with demand.

A multi-chain scaling roadmap

Block.one also outlined inter-blockchain communication as its preferred route to scale. The proposal envisioned multiple chains operated by the same elected producers, using the same native token while maintaining separate memory pools. Applications could pass authenticated messages between those chains through proofs.

Header-only validation was presented as a foundation for both light clients and cross-chain messaging. Block.one also described a revised last-irreversible-block algorithm intended to prevent conflicting finality unless more than one-third of producers behaved Byzantine. These were protocol claims and design objectives on May 4, not independently verified guarantees.

The company projected that a simplified, single-threaded EOSIO implementation might eventually process 5,000 to 10,000 transactions per second. No public benchmark accompanying the May 4 announcement established that range, so it should be read as a forward-looking estimate.

Governance reached into the protocol

Dawn 4.0’s blueprint retained 21 active block producers and proposed dividing annual token inflation between per-block and vote-based compensation. It also introduced vote decay, requiring voters to refresh their choices to preserve maximum influence.

Exchange integration received explicit attention through a deposit-monitoring tutorial and example software. That was institutionally important: EOSIO was being designed not only as a developer platform but also around the operational requirements of custodial trading venues.

Block.one emphasized that it was publishing open-source software and would not itself launch a public blockchain. Adoption, configuration and operation would remain the responsibility of third parties.

Later context

On May 11, 2018, Block.one formally released Dawn 4.0 and modified parts of the May 4 blueprint, including additional RAM-market guidance and a return to subjective CPU billing. That later release confirms why the May 4 record is best understood as a consequential design preview rather than final deployed software.