The Swarm
One video delivered by its own audience. Change who is watching and when, and see how much peers carry, how much needs an always-on node and where it stalls.
One video delivered by its own audience. Change who is watching and when, and see how much peers carry, how much needs an always-on node and where it stalls.
Experiment 04 · Distribution
An origin, dedicated seeders or blind peers that never go offline.
Assumptions, not measurements: with few phones, 50% of viewers can upload; with most on phones, 12%. The video comes in 3 qualities, and from the catalogue about 30 of 100,000 viewers watch at any one moment.
One video, delivered by its own audience. Choose who is watching, and when.
100,000 people press play on the same video at the same moment.
Nobody holds anything yet, so the first segments come from always-on nodes.
About 50% of viewers can accept connections and upload; the rest, on phones or hard networks, mostly take.
Peers pass segments to each other. In this model they carry 68% of what viewers watch.
Always-on nodes serve the other 32%. Nobody stalls, but the origin never goes away.
Every post and disc on the plinth is a viewer. Tall ceramic posts can accept connections and upload; short discs, mostly phones and viewers on hard networks, mostly take. At the back left stands the always-on node: an origin, a dedicated seeder or a blind peer. The viewer at the front is the one we follow. It lights while its video plays.
Every number here is an assumption you can see, not a measurement: how many viewers can upload, how much each can give, how many watch at once from the catalogue and the cost of the live edge. Real swarms add piece selection, churn over time, network distance and caching. The shape is the point: size helps, phones and the long tail hurt, and availability needs something that never sleeps.
This experiment accompanies What Happens When a P2P Video Has 100,000 Viewers?