<aside> ⏰ On hold because: We have solid theoretical and empirical justification for the algorithm. The pull request is waiting for review. We need the sign-off of the IPFS stewards for inclusion. On hold resolution: We got feedback from the IPFS stewards regarding the implementation to continue the work.
</aside>
Speeding Up the Provide Process by an Order of Magnitude
<aside> 🛠 Tasks towards Research and Deployment
</aside>
Material: Proposal | GitHub Discussion | GitHub Repo | Google Doc (obsolete) | Slides | Video | go-libp2p-kad-dht Fork | Pull Request | Overleaf
For each peer, that we come across during content publication, calculate the probability of an even closer one. If the likelihood is low, just store the provider record at that peer optimistically.
There are two things to be estimated
In this section, we will describe how libp2p DHT nodes can locally infer the global network size from routing table maintenance operations that they are doing anyways.
Literature Review