Protocol-Oriented Comparative Analysis of Video Streaming over DASH, TCP, UDP, and BPP
by HAVADER Editör Ekibi
When your internet slows down while watching a video, would you rather the picture quality drops, or the video freezes and buffers? That simple question actually sums up one of video streaming's most fundamental engineering dilemmas. This study comparatively examines that dilemma across four different protocols.
Video streaming performance is strongly shaped by transport-layer behavior under bandwidth constraints. This study's goal was to compare four video transmission approaches — DASH streaming, TCP-based transmission, UDP-based transmission, and the newer Big Packet Protocol (BPP) with trimming support — within a controlled experimental framework, examining delivery continuity, quality variability, and low-latency playback behavior.
The results showed DASH achieving the most consistent playback continuity in moderate, time-varying conditions through segment-level adaptation. BPP, a newer technology, offered continuity close to DASH's in dynamic scenarios by enabling "graceful degradation" through trimming. UDP proved highly sensitive to bandwidth limitation and data loss, leading to unstable reception and persistent interruptions, while TCP preserved constant quality thanks to retransmissions but caused the longest stall durations — making it less suited to low-latency streaming goals.
What this study contributes is clearly showing there's no single "best" protocol — each is optimized for a different priority. An everyday analogy: it's similar to a delivery company choosing between delivering a package in perfect condition but late (like TCP) versus slightly damaged but on time (like UDP) — DASH and BPP find a clever middle ground between the two, sacrificing a bit of quality when needed to preserve continuity.
In the end, this research gives engineers building live-streaming, video-conferencing, or online education platforms a practical guide to which protocol to prefer in which scenario — whether the bottleneck is fixed or dynamic.
Video streaming performance is strongly shaped by transport-layer behavior under bandwidth constraints. This study's goal was to compare four video transmission approaches — DASH streaming, TCP-based transmission, UDP-based transmission, and the newer Big Packet Protocol (BPP) with trimming support — within a controlled experimental framework, examining delivery continuity, quality variability, and low-latency playback behavior.
The results showed DASH achieving the most consistent playback continuity in moderate, time-varying conditions through segment-level adaptation. BPP, a newer technology, offered continuity close to DASH's in dynamic scenarios by enabling "graceful degradation" through trimming. UDP proved highly sensitive to bandwidth limitation and data loss, leading to unstable reception and persistent interruptions, while TCP preserved constant quality thanks to retransmissions but caused the longest stall durations — making it less suited to low-latency streaming goals.
What this study contributes is clearly showing there's no single "best" protocol — each is optimized for a different priority. An everyday analogy: it's similar to a delivery company choosing between delivering a package in perfect condition but late (like TCP) versus slightly damaged but on time (like UDP) — DASH and BPP find a clever middle ground between the two, sacrificing a bit of quality when needed to preserve continuity.
In the end, this research gives engineers building live-streaming, video-conferencing, or online education platforms a practical guide to which protocol to prefer in which scenario — whether the bottleneck is fixed or dynamic.