An aria2c based speedtest

This shell speed test will benchmark 8Gbps internet connections. It will :

Project Direction and Technology

This project aims to create a robust command-line internet speed test tool. The script will leverage the power of aria2c to perform downloads from a list of test servers, providing insights into internet connection performance.

Technological Choices

• Core Scripting: The primary script (src/speedtest.sh) is written in Bash, ensuring wide compatibility and access to powerful shell utilities.
• Download Engine: aria2c is used for its multi-connection and multi-segment download capabilities, which are essential for accurately benchmarking high-speed connections. The script utilizes options such as --min-split-size, --max-concurrent-downloads, --split, and --max-connection-per-server to optimize downloads.
• Configuration: Test server URLs are managed externally in the test-servers-list.txt file, allowing for easy updates and customization.
• Logging: Detailed logs of script operations and aria2c output are stored in ./debug.log.
• Output & Real-time Statistics: Currently, aria2c provides real-time download progress and speed information directly to the console during its operation. The README.md mentions a future goal of graphing real-time stats at 10fps and displaying values in both Gbps and MBps. Achieving highly customized real-time graphing and specific unit conversions might require additional tools or libraries (e.g., ncurses for terminal UI, bc for calculations, or specialized CLI graphing tools compatible with Fedora 42). This will be explored in later development phases.
• Concurrency: The script currently processes servers from test-servers-list.txt sequentially. The README.md mentions threading for simultaneous downloads. True parallel execution of multiple aria2c instances against different servers will be a key feature to implement in future phases to maximize test throughput.

• Be based around the following aria2c commmand options :
  aria2c --min-split-size=1M --max-concurrent-downloads=16 --split=16 --max-connection-per-server=16 --dir=/dev --out=null --quiet=true $(TEST_URL_BIG_FILE)

• Download directly to /dev/null so storage is not a bottleneck

• Be threaded, so it can launch this command simultaneously for as many test URLs provided in test-servers-list.txt, for instance :

http://appliwave.testdebit.info/1G.iso
http://speedtest.milkywan.fr/files/1G.iso
http://scaleway.testdebit.info/1G.iso

• Use the most modern and efficient shell libraries available on Fedora 42

• Log to ./debug.log

• Graph real time stats refresh at 10 fps

• Display realtime bandwidth values in both Gbps and in MBps.

• Have a cmdline option should activate looping : re-starting individual downloads when they end