Re: POP3 "resume download" extension
Keith Moore <[email protected]>
| Newsgroups | gmane.ietf.pop3ext |
|---|---|
| Message-ID | <[email protected]> |
> Thinking about pipelining, though, how does this sound: POP3 is normally a > half-duplex-like dialog. Suppose we allow an allowed a received command > sequence (like maybe "!NEXT") to queue an interrupt of the download at > the next chunk. If you're suggesting that servers do out-of-order command processing, I don't think that's the way to go. The goal of pipelining chunk requests is to get maximum download speed while still having an early abort capability. Say a client is trying to download a 1MB message in 100K chunks. If it has to get the response from each chunk request before asking for the next the next chunk, there's a round-trip delay after every chunk which could take as long as downloading 100K of data. So having partial retrieve could make the overall download take twice as long. OTOH, if the client can queue ahead several PRET requests then it can make sure that there is always a request outstanding in the server's queue...so there is no waiting. but it can still abort a download within the time it takes for the server to deal with the number of queued-ahead requests ... which is much better than waiting for the entire message and (if tuned right) also better than closing and reopening the connection. maybe what is needed are two "special cookies" for PRET: cookie # means "start at beginning of message" cookie @ means "continue transfer following the last PRET" however, this adds some state to the server. what happens if a client is displaying multiple messages from the same server in different windows, and the user keeps scrolling the windows around? the client (having initially only downloaded enough of each message to display the first page or two of each) might then want to interleave PRETs from one message with PRETs from another. Keith Keith