BlockBeats News, September 25th. Ethereum co-founder Vitalik Buterin mentioned on social media that the Fusaka upgrade will address the current data availability issue, with security always being the top priority. The core feature of the Fusaka upgrade, PeerDAS, is attempting to achieve an unprecedented breakthrough: allowing a running blockchain to operate without any single node downloading the complete data. PeerDAS operates by having each node request only a small amount of “data blocks,” ensuring that over 50% of the data blocks are probabilistically available. If more than half of the data blocks are available, nodes can theoretically download these data blocks and use erasure coding to recover the remaining parts. In the initial version, there are still two scenarios where the complete block data set needs to be stored:
(i) Initial broadcast stage;
(ii) Reconstruction phase when the publisher only releases 50%≤p
However, these roles do not require trust: only one honest actor needs to be present to complete the task, and even with 100 malicious nodes, the protocol will automatically circumvent them. Different nodes can perform this task for different blocks. In the future, cell-level message passing and distributed block building will achieve the decentralization of these two functions. This is all new technology, and even after years of development, the core developers maintain a highly cautious testing attitude, which is wise. This is also why the number of data blocks will increase conservatively at first and then gradually ramp up. However, this is crucial for L2 scaling (ultimately will also be used for L1 scaling—when the L1 Gas limit is raised to require L1 execution data to be stored in data blocks).
