Module 0 · Breaking In
Why self-hosting is worth it instead of buying cloud services
- EP01
The pain of interactive video games
Why traditional streaming protocols can't handle heavy interaction.
- EP02
The self-hosted CDN answer: PPCDN architecture overview
Origin/Edge/ppcenter three-tier architecture explained in one go.
- EP03
Content sovereignty: control it yourself, the ultimate moat
How self-hosting avoids takedowns/throttling and keeps data in-country.
Module 1 · Transport & Encoding Basics
The on-ramp for junior/mid-level developers
- EP04
Transport tech showdown: RTMP / SRT / WHIP
A direct comparison from principles to latency behavior.
- EP05
How Simulcast works and what it's for
One publish stream, multiple bitrates, server delivers on demand.
- EP06
Publish/play security: signing, tokens, and anti-hotlinking
How HMAC signing and short-lived credentials work in practice.
Module 2 · Weak Networks, Quality & Observability
The technical core of the series, the heaviest chapter
- EP07
How to strengthen resilience on weak networks
Upstream QoS, congestion detection, and priority isolation.
- EP08
Video instrumentation: measuring real end-to-end latency
NTP clock sync + SEI/OBU markers to truly measure every frame's delay.
- EP09
Latency stability matters more than average latency
Why jitter affects experience more than the mean.
- EP10
How to quantify video quality
Turning "it looks choppy" into a set of comparable numbers.
- EP11
What Netflix's seamless switching teaches us
Making bitrate/path switches feel invisible to users.
- EP12
Real-world test: latency stability vs. major cloud providers
Data over talk, a real head-to-head comparison.
Module 3 · Advanced Architecture
Filling in the resource-isolation and elastic-scheduling thread
- EP13
P2P direct connect: faster than CDN for the last mile
P2P can beat going through Edge, and it's not just for 1v1.
- EP14
Node pool management: giving every app its own Origin / Edge
Nodes declare a pool, apps bind to one, and scheduling picks the least-busy node within it.
- EP15
The relationship between horizontal scaling and low latency
More machines buy capacity, not a lower latency floor.
- EP16
Handling sudden bursts of high concurrency
Three layers of elasticity — P2P, on-demand origin pull, quality degrade — plus manual scaling.
Module 4 · Practice & Verification
Delivering on "compile, deploy, test"
- EP17
Zero to deployed: run your own CDN in 15 minutes
One compose file brings up four components plus a smoke test.
- EP18
Recording & playback: adding VOD to live streaming
Two-phase split, fMP4 segments, async upload, presigned playback.
- EP19
Load-testing methodology: from a single box to global load
Layered testing: measured, modeled, unverified — don't substitute one for another.
Module 5 · Commercial Wrap-up
Doing the math for a CTO audience
- EP20
Is a 30% bandwidth saving realistic? A real-data review
Saving traffic vs saving egress: measured, modeled, and unverified, laid out.