In my understanding, in other P2P networks like BitTorrent, they use supernodes to help each node find peers to download data. Does filecoin implement the same system as BitTorrent?
To my knowledge, Filecoin will sit on top of ipfs, and therefore it will use the IPFS to find the nodes and download your data.
As I understand it, the whole point of Proof of Replication is that Filecoin “miners” [storers maybe more appropriate?] don’t get paid unless they can prove with some frequency that the data you want is indeed available on demand, and at a reliable latency. You can always improve your odds of retrieving data fast by making multiple Filecoin deals for redundant storage (this is similar in effect to using RAID 1E). You may also want to make some explicitly higher latency Filecoin backup storage deals, which are probably on reliable storage just farther away, so those backup copies can escape any large geographic area catastrophic events. These deals would probably need to be structured on some [smart] legal contract outside the Filecoin network directly, possibly utilizing IPLD to link between your countersigned contracts and Filecoin transactions.