paulbms (paul-bms)

For those not already aware, Nick from Celestia shared a well-written thread on data availability (https://x.com/nickwh8te/status/1656321650475044865). There’s no need to break what’s not broken or reinvent the wheel, but there’s something missing from his buffet analogy that being: inspecting the buffet or in our example the tomatoes on the buffet. Data Availability ≠ Data Storage ≠ Data Validity 🥫 #DataStorage is like canned food. 🍱 #DataAvailability is an all-you-can-eat buffet. 🕵️‍♀️ #DataValidity is the quality inspector. In Nick's explanation, "Data storage solves the need to preserve data for future retrieval, similar to canning food. It's sealed and stored for later consumption." This analogy emphasizes the ease of accessing stored data multiple times without additional charges or data loss. But here's the problem: a can is not transparent.

• 1 reply
• 2 recasts
• 4 reactions

I love to see more decentralized storage solutions popping up, but here is the thing: NO, YOU CAN'T CREATE A PUBLIC GOOD DATASET WITHOUT A REAL, DECENTRALIZED VALIDATION LAYER. NONE OF THESE SOLUTIONS CAN DELIVER THAT. Here is why ⤵️

• 2 replies
• 0 recasts
• 2 reactions

Data availability layers don't guarantee perpetual access to downloadable data; they are only available for a limited duration. Sustaining this data requires running an archival node since #DPoS or #PoS networks lack incentives for such retention. Consequently, this often leads to centralization, creating a singular point of failure along with risks of data inconsistency and corruption. Essentially, foundations, validators, and third-party providers typically operate these archival nodes. While some foundations might provide free access, it's not consistently available. Users usually incur expenses to query data from these nodes, resulting in trust concerns, high costs, and substantial time investments for syncing if they opt for a local node setup. Hence, the significance of storing historical chain data in incentivized environments such as #Arweave, #EthStorage, #BNBGreenfield, or #Filecoin. Plus, employing tools like #LighthouseWeb3 further enhances this storage approach.

• 1 reply
• 0 recasts
• 1 reaction

2/ Even when multiple data sources upload to a decentralized storage solution, there’s no guarantee you’re retrieving the correct data. (all examples are with Arweave, but work for any solution)

• 1 reply
• 0 recasts
• 0 reactions