async-range-downloader
Utilizes the Range HTTP header to download a file with multiple connections in parallel, bypassing cheap bandwidth throttling
The in parallel running workers are lightweight coroutines built on ktor's HttpClient and coroutine wrappers around java.nio
Caveats
- Server must support the Range header
- Server side throttling must be on a per-connection-basis, not IP address
Usage
Usage: ard [options...] url [file]
Arguments:
url -> URL of file to download { String }
file -> Path of the output file (optional) { String }
Options:
--workers, -w [10] -> Number of async workers { Int }
--delay, -d [1] -> Delay between opening requests in seconds { Int }
--help, -h -> Usage info
If a filehost for example only allowed free-tier users to download at 1MB/s but your bandwidth allows for 10MB/s, you'd use 10 workers
TODO
- No error handling whatsoever at this time, should at least care for response codes
- Adding options for custom headers, cookies, or even auth
- Proxy support to bypass IP based throttling
