Live data from Hacker News

QUIC File Transfer Service, a CLI and srv for transferring large files

github.com

11–20 of 20 posts

Re: QUIC File Transfer Service, a CLI and srv for transferring large files

#11
Looks like the last time the code itself was touched was about seven months ago. And there are no open issues. I did spot the following in the README.md file: ``` An optional MD5 checksum can be calculated as well for the transferred file to verify that the content is the same as the source file. The server provides its own MD5 for comparison once the file is written. However, this would only be an additional level of redundancy as quic has a reliability guarantee already built into the protocol. ``` Could we make this SHA-256 or something else a little more modern and robust than MD5?

Re: QUIC File Transfer Service, a CLI and srv for transferring large files

#14

Is it time to move QUIC past UDP yet? Does QUIC deserve it's own Protocol number? There's only one byte in IP for protocol number, so it makes sense to limit the number of applications, but I believe QUIC has reached that level. https://www.iana.org/assignments/protocol-numbers/protocol-n...

The problem is that most routers don't support anything other than UDP and TCP.

Re: QUIC File Transfer Service, a CLI and srv for transferring large files

#15

Weird to be comparing vs rsync. Rsync is notorious for maximizing latency as it has no pipelining. Better baseline would be transferring data via ssh + tar. Or serving a directory over http and concurrently downloading a file with axel

I would be more interested in seeing throughput vs rsync or tar+ssh (or scp, really).

Re: QUIC File Transfer Service, a CLI and srv for transferring large files

#16

Looks like the last time the code itself was touched was about seven months ago. And there are no open issues. I did spot the following in the README.md file: ``` An optional MD5 checksum can be calculated as well for the transferred file to verify that the content is the same as the source file. The server provides its own MD5 for comparison once the file is written. However, this would only be an additional level o…

Hey man, had no reason to work on the code further, was a one man experimentation with no incentive to work on it further. The use of MD5 was not meant for security guarantees but instead as a way to ensure the integrity of content. There would be no reason to use SHA-256 in this case, not worried about security since that would be handled through use of SSL/TLS between server and client, which is already available.

Re: QUIC File Transfer Service, a CLI and srv for transferring large files

#17

Well.. I'm not a go programmer. Just wanted the usual download/install and syntax to use the tool. Nowhere to be found. I guess I'll just open a ticket there.

Hey, you can already do this. It is designed to be portable and usage is described in the cmd folder in separate markdown file. Can be used in docker as well, with server and client usage described.

Re: QUIC File Transfer Service, a CLI and srv for transferring large files

#18

Weird to be comparing vs rsync. Rsync is notorious for maximizing latency as it has no pipelining. Better baseline would be transferring data via ssh + tar. Or serving a directory over http and concurrently downloading a file with axel

Sure it may be weird, but my company utilized rsync for almost everything, and we were transferring files that were 100+Gb on the regular so this was actually a great tool to compare against since we were already using it for large file transfers.

Re: QUIC File Transfer Service, a CLI and srv for transferring large files

#20
Hey, not sure how this came up but I’m the original author of the repository…thanks for taking a look at my code and thinking it was cool enough to post on here. If you have any questions let me know, or create an issue on the repository and I’ll try to take a look at it.
Post reply on HN