IPFS + (Arweave or Filecoin) behind it seems like the best decentralized web stack I've seen yet. I'm not sure how Orbit holds up.
Peer-to-Peer Databases for the Decentralized Web
11–20 of 84 posts
Re: Peer-to-Peer Databases for the Decentralized Web
#12I wonder if it would be possible to build a decentralized search engine with a decentralized crawl on top of it.
I've been wondering the same thing too! Maybe if there is a way to share indexes, or some way to put out a call for searches through the DHT, and then the requestor aggregates the results.
A couple examples:
Web crawler -- seeded through a curated or personal bookmark list.
Knowledge base -- the permaculture community have amassed a knowledge base on growing plants: USDA hardiness zones, heat zones, humidity, soil requirements, companion planting relationships, etc. That's something I can see people adding on to, particularly with non-standardized homegrown cultivars and how well they do in their local area.
Re: Peer-to-Peer Databases for the Decentralized Web
#13This project may be dead There has been no activity in over a year,and the sponsoring company has no presence in several years.
Gun DB is a good replacement.
Re: Peer-to-Peer Databases for the Decentralized Web
#14Earlier quoted context omitted.
What do you think decentralized means? It means that hosting it isn't the responsibility of a single party. If I put something online, and people think it's interesting and pin it. Even if I go down, the pinned content remains up. Decentralized doesn't magically solve hosting availability. It just removes responsibility from a single party.
I would say there are very few IPFS sites where hosting doesn't still end up being the primary responsibility of a single party. That was the point I was trying to make. If you have to pay someone else to pin it and keep giving them funds then it is still your responsibility. If you have a website that you want to scale and load fast for all your users on IPFS then it will still end up being your responsibility. Ther…
There is no technology possible that removes said responsibility. This is how peer to peer systems work. Your criticism is something that no peer to peer system claims to solve.
Re: Peer-to-Peer Databases for the Decentralized Web
#15I don't see how people consider this and IPFS decentralized. You still need peers to subscribe and pin, and services like Pinata aren't cheap and still centralized. Is there an example of websites that would continue to run indefinitely if there wasn't a centralization of services with IPFS? I just don't think it is likely unless you start distributing your app that bundles IPFS, so that your users end up hosting it…
I'm looking at integrating IPFS with sandstorm (sandstorm.io). The sandstorm (and other homelab/homeserver) community have an interest in self-hosting. IPFS allows people that are interested in seeing something stored to pin it themselves. But yes: bundling IPFS with the app is pretty much what would need to happen. As they are now, neither the IPFS ecosystem or sandstorm are really decentralized -- or maybe I should…
Curious about the homelab/homeserver community. Are there any sites with further information?
Re: Peer-to-Peer Databases for the Decentralized Web
#16I wonder if it would be possible to build a decentralized search engine with a decentralized crawl on top of it.
I've been wondering the same thing too! Maybe if there is a way to share indexes, or some way to put out a call for searches through the DHT, and then the requestor aggregates the results. A couple examples: Web crawler -- seeded through a curated or personal bookmark list. Knowledge base -- the permaculture community have amassed a knowledge base on growing plants: USDA hardiness zones, heat zones, humidity, soil re…
Limewire had an interesting protocol for exactly this back in the day, over its Gnutella network. I don't know if it scales to search engine sizes (most things don't, especially distributed) but there's at least some prior production grade work in this space.
Edit: having just double checked it looks like Gnutella predates limewire; that's just the avenue through which I came across it.
Re: Peer-to-Peer Databases for the Decentralized Web
#17This project may be dead There has been no activity in over a year,and the sponsoring company has no presence in several years.
It's not dead, just in a bit of a lull due to funding, which is hopefully coming soon. We have a very active Gitter and OpenCollective.
Source: I'm one of the core maintainers
Re: Peer-to-Peer Databases for the Decentralized Web
#18Why do i always feel things like this and ipfs cone back to blockchain? I think the blockchain is a very interesting tool in a limited capacity. Orbitdb and ipfs seem to capitalize on their relationship to it.
OrbitDB doesn't rely on a blockchain at all
Re: Peer-to-Peer Databases for the Decentralized Web
#19Earlier quoted context omitted.
I would say there are very few IPFS sites where hosting doesn't still end up being the primary responsibility of a single party. That was the point I was trying to make. If you have to pay someone else to pin it and keep giving them funds then it is still your responsibility. If you have a website that you want to scale and load fast for all your users on IPFS then it will still end up being your responsibility. Ther…
There is no technology possible that removes said responsibility. This is how peer to peer systems work. Your criticism is something that no peer to peer system claims to solve.
Actually, I recommend you check out Freenet [0], a project that's 20 years old. In fact, the Wikipedia page explicitly states:
> Information flow in Freenet is different from networks like eMule or BitTorrent; in Freenet:
> 1. A user wishing to share a file or update a freesite "inserts" the file "to the network" > 2. After "insertion" is finished, the publishing node is free to shut down, because the file is stored in the network. It will remain available for other users whether or not the original publishing node is online. No single node is responsible for the content; instead, it is replicated to many different nodes.
Re: Peer-to-Peer Databases for the Decentralized Web
#20I don't see how people consider this and IPFS decentralized. You still need peers to subscribe and pin, and services like Pinata aren't cheap and still centralized. Is there an example of websites that would continue to run indefinitely if there wasn't a centralization of services with IPFS? I just don't think it is likely unless you start distributing your app that bundles IPFS, so that your users end up hosting it…
Swarm persistence is hard, if not close to impossible. Better that multiple parties pin the data than just one though. It's a step in the right direction.