Need Assistance Sussing Out Consensus-Based Storage Use Case

I am contemplating building a project akin to devdocs.io (if you are unfamiliar, see here) that caches client-side in-browser for fast lookup. I am looking to use WebTorrent or IPFS and am intrigued by the in-browser fetch possibility. My struggle is that I want the scraping that would traditionally be done by a single party and put at an IPFS or IPNS hash to be done by multiple parties. Then I would like the client to somehow use a consensus model to determine the most popular shared hash to prevent spoofing and centralization. I don’t want the clients to have to build the hash themselves, only the people running the daemons.

Any suggestions on how to do this amongst IPFS peers? I guess this sounds similar to a blockchain or merkle DAG. I’m really trying to avoid a central entity blessing a given hash. I’m beginning to think this isn’t really possible without a web-of-trust like trust the client has in someone’s peer key on an IPNS hash. Any other projects out there using IPFS w/ consensus models and adversary protection? Can I use the DHT directly to build consensus maybe?

Yes, in this case, consensus is a bit difficult without a web of trust (or some permissioned system with voting).

1 Like