Computer Networks · Assignment 1 · ABR algorithm comparison
idle
Live QoE metrics
Quality switches
0
Rebuffering events
0
Total rebuffer time
0.00s
Startup delay
–s
Current bitrate
–kbps
Avg bitrate (time-wtd)
–kbps
Buffer level
0.0s
Est. throughput
–kbps
Playback / representation
–
Event log
All dash.js events are captured to the log regardless of what is shown here.
Stream configuration
Your server’s IP and port — not this machine. Saved between reloads.
The 75M/180M/300M renditions are H.264 level 6.2
(2880p/4320p). Browsers have no hardware decode path for these, so
software decode stalls and playback dies mid-stream. Capping keeps the
run valid; under 4G/3G throttling these rungs are never selected anyway.
Throttling is set in DevTools; this is just how the run is named.
Include the network in the label (Fast_4G,
Slow_4G, 3G, 5G) and it is
parsed into the log automatically, so the plots group correctly.
Logging
Leave blank to log in-browser only. See log_server.js.
Metrics are also saved automatically when playback ends.
Manual quality lock
Forces one rendition, bypassing ABR. Use it once before the experiments to
confirm the top 8K renditions actually decode in this browser.
Throttling reference
DevTools → Network → Throttling
Profile
Down
Up
RTT
Fast 4G
4.0 Mbps
3.0
20 ms
Slow 4G
1.6 Mbps
0.75
150 ms
3G
750 kbps
0.25
100 ms
Chrome’s built-in presets. If your build only lists “Slow/Fast 3G”,
add custom profiles with these values and state them in your report.
Set throttling before pressing Load, and hard-reload
(Ctrl‑Shift‑R) between runs so nothing is served from cache.