3 ms·
This brings to mind one of the issues I have with the w3 model: It eliminates efficiencies of scale. All of the processing to render a website is delegated to c
by falcolas 5y ago
This brings to mind one of the issues I have with the w3 model: It eliminates efficiencies of scale. All of the processing to render a website is delegated to customers (many of whom are on cell phones with terrible bandwidth). You can't do any sort of caching or any sort of optimizations to make a website run faster. You couldn't do it even if you needed to for your customers.
No caching. No QUIC. No parallel pipelines (you need the whole blockchain locally to begin rendering), no deletion of old content. No CDNs. No edge distribution of secured media (Netflix movies, et.al.).
Everybody must (to be secure) verify all the entries of the entire blockchain as a boot-up step. At least once per load, viewers must traverse the block chain to build the latest state for a specific web page.
It gets worse if you optimize for the above actions by creating and maintaining one block chain per individual or small groups of w3 sites. Frankly, I find this idea to be completely unrealistic. One can look at existing P2P and see how seeds vanish without traditional websites (and their trackers - hey, databases!) to back them up.
Just imagine what Reddit's blockchain (size, transaction time, parsing time, growth) would look like.
- birracerveza 5y agoThat's not how it works. The users' clients receive already verified data, the "verification" is done by the network nodes that provide it. The network can cryptographically ensure that the data is not tampered with, so that every node in the network returns verified, secure data, regardless of who you're retrieving it from. The user does not have to download the blockchain. It would be the same as needing to download a database, it makes no sense. It all happens in the backend, all the client needs is a simple API call.