Network Upgrade Targets Quantum Security by 2029

The blockchain network currently ranked second by market capitalization is mapping out significant structural changes. Recently, a roadmap for the upcoming Hegotá upgrade was published, targeting a fully quantum-resistant base layer by 2029. This long-term evolution will prioritize EIP-8141, focusing on native account abstraction as a critical first step.

Simultaneously, researchers are exploring near-term methods to reduce the computing burden associated with layer-2 rollups. A September 3 report by researcher Csaba Kiraly details a prototype design that reported an 11 to 18 times reduction in estimated reconstruction work during simulations involving 1,000 nodes.

Streamlining Data Recovery

The network utilizes a system called PeerDAS to verify the availability of data blobs, which carry information used by rollups. Under this framework, nodes are not required to download the entirety of the data. High-custody nodes maintain at least 64 of the 128 data columns, which is enough to rebuild any missing blob data. Meanwhile, supernodes hold all 128 columns.

In the existing model, multiple high-custody nodes often end up duplicating the exact same data reconstruction tasks. The baseline PeerDAS system does include randomized waiting periods and checks designed to suppress duplicate efforts. However, Kiraly’s reduced design aims to lower duplication further by assigning specific blobs to designated nodes first. Other network participants can then receive the recovered information through existing column-distribution channels instead of immediately executing the rebuilding process themselves.

The prototype measured accumulated computing work across the simulated network rather than elapsed recovery time, utilizing a baseline cost of 162 milliseconds per blob recovery on a Ryzen 9 8945HS processor. Measurements specifically excluded transaction speeds and fee savings.

In a simulated configuration featuring four blobs, no withheld columns, and a 10% share of supernodes, the network-wide reconstruction cost dropped from 48.6 CPU-seconds under the standard PeerDAS model down to 2.75 CPU-seconds. When the supernode share was increased to 20%, the estimated computing cost fell from 91 CPU-seconds to 6.6 CPU-seconds.

Testing Limitations and Future Integration

This reduced prototype is positioned as a potential preliminary step toward a broader networking proposal known as RowDAS. Specified in the draft EIP-8371, full RowDAS would introduce row channels that allow smaller nodes to pool their data and reconstruct information collectively once a recovery threshold is met. By contrast, the reduced design operates without these new row channels, and relies on high-custody nodes to maintain a delayed recovery role as a PeerDAS-style backstop for missing information.

Because it lacks the row layer, the prototype cannot provide the additional resilience of the complete EIP-8371 proposal, which currently leaves blob limits unchanged. The proposed split between duty assignment and row networking has not yet been incorporated into the draft text.

Furthermore, the measurements outlined in the September 3 report remain restricted to in-process, simulated networks utilizing real cryptography. Devnet results have not yet been reported, and the full 128-row-subnet configuration relies on extrapolations from tests with smaller subnet counts. Larger simulations and real-network tests are still ahead.