Peer to peer? I guess I am getting old. I can remember when peer to peer meant that one host would communicate directly to the other host and not store the file transfer on a centralized system. Or that an application would not use a centraliozed third party for authentication.
Amen, brother! These were my assumptions as well when clicking the link. It would be more accurate to call this service peer-to-server-to-another-server-to-peer.
We only store the file on our central server if the receiving end doesn't have our app. If they have the app, we create a direct connection between both parties and transfer the file(s) directly without them ever hitting our servers.
That's a little disappointing, because it means your claim to have no file size limits can't really be true. What is the real limit for sending a file to someone who doesn't have the app installed? Would a 100TB transfer really work?
I don't mean to be snide; I honestly hoped what was going on was that your server was only brokering the transfer, not taking a full upload. In other words, only when the recipient began the file download would sendrail begin uploading the file from my computer. In this scenario, sendrail would merely be reflecting a stream, not storing anything.
So tell me if I understand correctly: If I share a document with someone who DOESN'T have sendrail, they will get the version of the file at the moment I send it. But if I share a document with someone who DOES have sendrail, they will get the version of the file at the moment they download it? In other words - peer to peer downloads get any file changes since the file was sent, but non peer to peer do not?
Curiosity question: what happens if I disconnect my machine halfway through a direct transfer? Do you support continued download/upload (or multipart)?
I guess in the case of a connection to your server, presumably the receiving party would never know that you sent them something (until a retry upload happened).
It's just that in the case of a direct p2p transaction, it's quite likely that the downloading user's downstream bandwidth is low, such that the chance of a failure happening over the entire course of the longer transfer is higher.
