Cut Delivery Cost
Rilla lowers delivery cost when viewers relay video using their upload bandwidth instead of relying only on CDN delivery.
Concurrency drives delivery cost
Section titled “Concurrency drives delivery cost”Live delivery cost rises with the same factors that make major events valuable: more viewers, longer sessions, higher bitrates, and compressed peak windows.
In a CDN-only model, each additional viewer and minute watched adds more CDN-served traffic. During premium live sports, that can create short, intense cost spikes where audience growth maps directly to transport spend and capacity exposure.
Rilla changes the cost model when peak audience growth no longer maps directly to CDN-served traffic.
Capacity planning may also account for infrastructure that is only required during the highest-demand moments. That makes event economics harder to predict before the actual audience arrives.
Deflection reduces CDN-served traffic
Section titled “Deflection reduces CDN-served traffic”Rilla reduces CDN-served traffic through CDN deflection: suitable live video traffic is delivered through Rilla’s peer-assisted layer instead of directly from CDN infrastructure.
Your CDN remains available. It continues to serve traffic that is not eligible for peer delivery and acts as the fallback when peer delivery cannot meet playback requirements.
The goal is not maximum deflection at any cost. The goal is commercially meaningful deflection while holding playback quality within agreed guardrails.
In suitable live-streaming conditions where ~30% of the audience can consistently contribute bandwidth, Rilla typically targets 60–90% CDN deflection while holding QoE guardrails steady. Results depend on audience overlap, eligible platforms, and rollout scope (CDN Deflection).
Event economics improve when traffic shifts
Section titled “Event economics improve when traffic shifts”For live and linear streaming, many viewers often watch the same segments at roughly the same time. That shared demand creates an opportunity for one CDN-served segment to support more than one viewer through peer relaying.
When Rilla safely deflects traffic, viewers can relay suitable segments using their upload bandwidth, so the amount of video data your CDN must serve directly can decrease. That can improve cost per event and delivery cost per viewer, especially during high-concurrency windows where CDN usage would otherwise rise sharply.
Actual savings depend on traffic shape, eligible platforms, CDN contracts, bitrate mix, rollout scope, and measured deflection. POC results validate production business cases before broader rollout.
Cost metrics to measure
Section titled “Cost metrics to measure”Cost validation connects delivery measurements to commercial inputs. These metrics are strongest when measured during representative live traffic and compared to a CDN-only control group.
| Metric | What it shows |
|---|---|
| CDN deflection rate | Share of suitable live video traffic delivered through Rilla instead of directly from CDN. |
| Peer-delivered GB | Volume delivered through the peer-assisted layer. |
| CDN-delivered GB | Remaining traffic served by CDN after deflection. |
| CDN cost per event | Event-level impact based on the CDN pricing model. |
| Delivery cost per viewer | Derived measure for whether delivery economics improve as concurrency increases. |
| QoE delta vs control | Whether cost reduction stayed within playback guardrails. |
Proving savings in a POC
Section titled “Proving savings in a POC”POC baselines define the comparison point for savings and keep the business case grounded in measured delivery behavior rather than assumed offload.
Select the evaluation scope
Choose a representative live stream, region, platform, or audience cohort.
Compare against a control
Measure Rilla-enabled viewers against CDN-only delivery under the same event conditions.
Measure delivery distribution
Track deflection, peer-delivered GB, CDN-delivered GB, and QoE deltas.
Map traffic reduction to cost
Apply the CDN pricing model to estimate event-level commercial impact.
Expand only after validation
Increase rollout when deflection and playback guardrails meet agreed thresholds.