Skip to content

Capacity on demand

Rilla unlocks capacity on demand as viewers join the same live stream and relay video to one another.

Live sports peaks are not smooth traffic curves. They concentrate around moments that matter: match starts, finals, kick-off windows, major plays, regional surges, breaking news, and viral moments inside a live stream.

Rilla adds peer-assisted capacity for these windows alongside your CDN, so viewers can deliver video segments to others watching the same stream.

CDNs are essential to live streaming, but peak events can force worst-case capacity planning. A platform may need to reserve or pay for capacity that is only used for a short event window, while still carrying the risk that actual demand exceeds forecast. Rilla does not remove the need for CDN planning; it adds audience-powered capacity that can reduce pressure on CDN-only delivery when live conditions support peer assistance.

Peak load usually combines several pressures at once:

PressureWhat changes during a peak event
ConcurrencyMore viewers need delivery at the same time.
Join-window compressionLarge audiences arrive before kick-off or during a major moment.
Regional concentrationTraffic can concentrate in specific regions, ISPs, or access networks.
Quality exposurePlayback failures are most visible during premium moments.

Traditional delivery capacity is usually provisioned before demand arrives. Rilla’s capacity emerges from participation.

When viewers join a stream, eligible devices can relay encrypted video segments to other viewers. This is most useful for live and linear streaming, where many viewers are watching the same content segments at roughly the same time.

Your CDN remains the primary delivery mechanism, and each viewer keeps a CDN connection available for instant fallback when peer delivery is unsuitable.

Real-time coordination absorbs concurrency

Section titled “Real-time coordination absorbs concurrency”

As concurrency increases, available peer capacity also increases. Rilla evaluates which viewers are suitable contributors based on reliability, capacity, playback state, and network conditions.

The AI Orchestrator coordinates peer suitability in real time. The Performance P2P SDK executes delivery decisions inside the player. If peer delivery cannot meet playback requirements, traffic falls back to your CDN.

This model absorbs peak demand while preserving the existing CDN, player, DRM, entitlement, and monitoring workflows.

Capacity validation uses real or representative high-concurrency conditions, not only theoretical network capacity. The strongest validation compares Rilla-enabled viewers against a CDN-only control group during the same live-event conditions.

Useful POC metrics include:

KPIWhy it matters
Peak concurrent Rilla-enabled usersShows the scale of the evaluated cohort.
CDN offload rateShows how much traffic Rilla absorbed during the event window.
P2P connection success rateShows whether viewers could form useful peer connections.
Rebuffer ratio delta vs controlConfirms added capacity did not harm playback.
Video startup time delta vs controlConfirms Rilla did not slow playback start.
Fallback frequencyShows how often CDN protection was required.