The bootstrap node is an IPFS node that other nodes can connect to in order to find other peers. Now all the bootstrap nodes are running by IPFS development team. Is it possible that all bootstrap nodes on IPFS network were blocked in certain area and render IPFS network un-accessible for users within the area?
Currently yes, but in the future its plan to do something like i2p.
A lot (at least ~3000) of peers we successfully connected to will be stored in a file and some randomly picked will be used the next time for bootstrap (and if that unsuccessful more peer can be used, that will take a lot of time until all peers run out). So bootstrap peers will be used only the first time, and again like i2p if they are down a friend can share you some peers of his peerstore OOB to get you started.
Yes and no.
You can share a list of known good nodes with the program, like the official binary is shared with the default bootstrap nodes.
There’s some risk of a netsplit involved, if you share a very limited amount of bootstrap nodes and you need to have a limited trust that they will not all controlled by the same authority.
How to easily save all current bootstrap nodes in a file, remove the default bootstrap nodes and import others from a file is documented here:
You can also get a list of current connections of a ipfs daemon, and save this list to a file:
This way you could simply build a large list of peers and randomly import some of them on each daemon start by hand or by scripting.
That’s basically what @Jorropo describes for the future.
1 Like