--- title: "WebRTC vs CDN Streaming: When Sub-Second Latency Is Worth It" description: WebRTC vs CDN-based HLS/DASH streaming compared on latency, scale, cost and reliability, with hybrid patterns for interactive live video. image: https://blog.blazingcdn.com/hubfs/Gemini-Blog/image-May-29-2026-08-41-15-5646-AM.png --- [![BlazingCDN](https://blog.blazingcdn.com/hubfs/Logo/blog-logo-w.png)](https://blog.blazingcdn.com?hsLang=en-us) [๐Ÿ“˜Learn โ–พ](https://blog.blazingcdn.com/cdn-learn?hsLang=en-us) [CDN Fundamentals](https://blog.blazingcdn.com/cdn-fundamentals?hsLang=en-us) [By Content Type](https://blog.blazingcdn.com/by-content-type?hsLang=en-us) [Advanced Concepts](https://blog.blazingcdn.com/advanced-concepts?hsLang=en-us) [Glossary](https://blog.blazingcdn.com/glossary?hsLang=en-us) [โšก Web Performance](https://blog.blazingcdn.com/web-performance?hsLang=en-us) [๐ŸŽฌVideo & Streaming โ–พ](https://blog.blazingcdn.com/video-streaming-cdn?hsLang=en-us) [๐Ÿ”ด Live Streaming](https://blog.blazingcdn.com/live-streaming?hsLang=en-us) [๐Ÿ“บ VOD & OTT](https://blog.blazingcdn.com/vod-ott?hsLang=en-us) [๐Ÿ’ฐ Bandwidth & Costs](https://blog.blazingcdn.com/bandwidth-costs?hsLang=en-us) [๐Ÿญ Other Industries โ–พ](https://blog.blazingcdn.com/cdn-industry-insights?hsLang=en-us) [๐Ÿ“บ Media & Broadcasting](https://blog.blazingcdn.com/media-broadcasting?hsLang=en-us) [๐Ÿ’พ Software & SaaS](https://blog.blazingcdn.com/software-saas?hsLang=en-us) [๐Ÿ—๏ธ DevOps & Cloud Infra](https://blog.blazingcdn.com/devops-cloud-infra?hsLang=en-us) [๐Ÿ“š AdTech & Advertising](https://blog.blazingcdn.com/adtech-advertising?hsLang=en-us) [๐ŸŽฎ Gaming & Esports](https://blog.blazingcdn.com/gaming-esports?hsLang=en-us) [๐Ÿ“ฑ Mobile Apps & Developers](https://blog.blazingcdn.com/mobile-apps-developers?hsLang=en-us) [๐Ÿค– AI & Machine Learning](https://blog.blazingcdn.com/ai-machine-learning?hsLang=en-us) [๐Ÿ›๏ธ Enterprise & Corporate](https://blog.blazingcdn.com/enterprise-corporate?hsLang=en-us) [๐Ÿ“š E-Learning & EdTech](https://blog.blazingcdn.com/e-learning-edtech?hsLang=en-us) [๐ŸŸ๏ธ Sports & Live Events](https://blog.blazingcdn.com/sports-live-events?hsLang=en-us) [๐Ÿ’ฐ Pricing & Costs โ–พ](https://blog.blazingcdn.com/cdn-pricing-and-cdn-costs?hsLang=en-us) [Provider Pricing](https://blog.blazingcdn.com/provider-pricing?hsLang=en-us) [Cost Optimization](https://blog.blazingcdn.com/cost-optimization?hsLang=en-us) [Decision Support](https://blog.blazingcdn.com/decision-support?hsLang=en-us) [โšกCompare โ–พ](https://blog.blazingcdn.com/cdn-comparison?hsLang=en-us) [Provider Comparisons](https://blog.blazingcdn.com/provider-comparisons?hsLang=en-us) [Strategy Comparisons](https://blog.blazingcdn.com/strategy-comparisons?hsLang=en-us) [Ratings & Benchmarks](https://blog.blazingcdn.com/cdn-ratings-and-benchmarks?hsLang=en-us) [๐Ÿ“Š Benchmarks](https://blog.blazingcdn.com/cdn-ratings-and-benchmarks?hsLang=en-us) [๐Ÿ”’ Security โ–พ](https://blog.blazingcdn.com/cdn-security?hsLang=en-us) [Attack Protection](https://blog.blazingcdn.com/attack-protection?hsLang=en-us) [Encryption & Access](https://blog.blazingcdn.com/encryption-access?hsLang=en-us) [Video & DRM Security](https://blog.blazingcdn.com/video-drm-security?hsLang=en-us) [๐Ÿ“ Case Studies](https://blog.blazingcdn.com/case-studies?hsLang=en-us) [๐Ÿ”Œ Integrations](https://blog.blazingcdn.com/integrations?hsLang=en-us) [๐Ÿ› ๏ธ Tools](https://blog.blazingcdn.com/cdn-tools?hsLang=en-us) [Get Started](https://blazingcdn.com/sign-up-contact-form/) [Learn](https://blog.blazingcdn.com/en-us/tag/learn) [Video - Video & Streaming](https://blog.blazingcdn.com/en-us/tag/video-video-streaming) [Video - Live Streaming](https://blog.blazingcdn.com/en-us/tag/video-live-streaming) [Learn - Advanced Concepts](https://blog.blazingcdn.com/en-us/tag/learn-advanced-concepts) # WebRTC vs CDN Streaming: When Sub-Second Latency Is Worth It BlazingCDN Sep 17, 2026, 7:10:24โ€ฏPM ![](https://blog.blazingcdn.com/hubfs/Gemini-Blog/image-May-29-2026-08-41-15-5646-AM.png) # WebRTC Streaming vs CDN in 2026: A Scaling Playbook A pure WebRTC streaming pipeline delivers glass-to-glass latency in the 50โ€“200ms range. A standard HLS chain delivers 6โ€“30 seconds. That four-order-of-magnitude gap is why interactive auctions, live trading dashboards, and real-money gaming run on WebRTC. But the same architecture that wins at 200ms collapses at scale: a mesh topology dies around 6โ€“8 peers, and even an SFU saturates well before you reach broadcast audiences. This article gives you the 2026 decision matrix for choosing WebRTC, CDN-delivered low-latency formats, or the hybrid that most production teams now ship โ€” plus the threshold values that tell you which one your workload needs. ![image-2](https://blog.blazingcdn.com/hs-fs/hubfs/Gemini%20INBlog%20Pictures/image-May-29-2026-08-41-33-0957-AM.png?width=1280&height=720&name=image-May-29-2026-08-41-33-0957-AM.png) ## Why traditional HLS and DASH fall short for real-time video Segmented delivery buffers by design. Even with LL-HLS partial segments and chunked-transfer CMAF, you are fighting the player's need to fill a jitter buffer before playback. As of 2026, well-tuned LL-HLS deployments land around 2โ€“5 seconds glass-to-glass in the field, not the sub-second the spec suggests on paper. That is excellent for a sports broadcast. It is unusable for a two-way medical consult or a synchronous auction where a 3-second lag means someone bid on stale data. The constraint is structural. HTTP segment delivery rides TCP, and TCP's retransmission and congestion control add unavoidable head-of-line delay under loss. WebRTC streaming sidesteps this by riding SRTP over UDP, trading guaranteed delivery for time. When a packet is late, real-time video would rather drop it than wait. ## WebRTC live streaming: built for sub-200ms, not for millions WebRTC gives you encrypted, plugin-free media with DTLS-SRTP, adaptive bitrate via the bandwidth estimator, and packet-loss concealment through NACK, RED, and FEC. For one-to-one and small-group sessions, nothing in 2026 beats it on latency. The problem is fan-out. - **Mesh:** Every peer encodes and uploads to every other peer. Practical ceiling is roughly 6โ€“8 participants before uplink bandwidth and CPU collapse. - **SFU (Selective Forwarding Unit):** Each publisher sends one upstream; the SFU forwards selectively. A single beefy SFU instance handles low thousands of subscribers per stream before egress and connection-state overhead force horizontal sharding. - **MCU:** Server-side mixing reduces client load but adds transcode latency and cost per session, rarely worth it for broadcast fan-out. Past the low thousands, you are no longer running a conferencing problem. You are running a distribution problem. That is where the CDN re-enters the picture. ## How does WebRTC work with a CDN for live streaming? The hybrid pattern that dominates in 2026 splits the audience by interaction need. Active participants โ€” the people who must talk back in under 200ms โ€” stay on WebRTC through an SFU tier. Passive viewers, who only need to watch with low latency, get a fan-out path optimized for scale. Two architectures matter here: - **WebRTC-to-WebRTC fan-out at the edge.** A cascade of SFUs distributes the same media tree across regions, with edge SFUs terminating viewer connections. This holds sub-second latency for very large audiences but is operationally heavy: every viewer is a stateful peer connection with ICE, DTLS, and continuous bandwidth estimation. - **WebRTC ingest, CDN egress in a low-latency format.** WebRTC handles the contribution leg, then a media server repackages into LL-HLS or low-latency CMAF for HTTP delivery through a CDN. You give up the sub-second floor for viewers (landing at 2โ€“5s) but gain near-infinite scale on commodity HTTP infrastructure and standard cache hierarchies. The honest tradeoff: stateful WebRTC fan-out buys you latency; stateless HTTP fan-out buys you scale and cost efficiency. Most teams run both legs and route per viewer role. ## WebRTC vs HLS for low-latency live streaming: the decision matrix Pick the path by the question your workload must answer, not by the technology you find interesting. This matrix reflects 2026 production realities. | Workload profile | Latency need | Concurrency | Best fit | | --- | --- | --- | --- | | Telemedicine, 1:1 consults | <200ms, bidirectional | 2โ€“4 peers | Pure WebRTC, mesh or single SFU | | Interactive auctions, betting | <500ms, synchronized | 10kโ€“500k | WebRTC fan-out via cascaded SFUs | | Live commerce, town halls | 1โ€“3s acceptable | 100kโ€“millions | WebRTC ingest + LL-HLS over CDN | | Sports, concerts, broadcast | 3โ€“8s acceptable | Millions | LL-HLS / LL-CMAF, pure CDN | ## Components of a real-time CDN architecture in 2026 The pieces have not changed in name, but their tuning has. - **STUN/TURN:** STUN resolves the public mapping; TURN relays when symmetric NAT or restrictive firewalls block direct paths. As of 2026, plan for 15โ€“25% of sessions to fall back to TURN relay, and budget egress accordingly โ€” relayed media is your most expensive byte. - **SFU tier:** Stateless-where-possible, regionally cascaded, with simulcast or SVC so the SFU drops layers rather than transcoding under congestion. - **Repackaging layer:** WebRTC-to-CMAF conversion for the HTTP egress path, where the CDN's cache hierarchy and origin shield do the heavy lifting. - **Telemetry:** Per-session RTT, jitter, loss, and freeze ratio, sampled at the receiver, not the server. The HTTP egress leg is where CDN economics decide your margin. For the LL-HLS fan-out path, [**BlazingCDN's media delivery infrastructure**](https://blazingcdn.com/solutions-for-media-companies/) offers stability and fault tolerance on par with Amazon CloudFront while running materially cheaper at volume โ€” pricing scales down to $0.002 per GB ($2 per TB) at the 2 PB tier, with 100% uptime and fast scaling under demand spikes. For a live-commerce event that spikes from 5k to 800k viewers in minutes, that combination of headroom and per-GB cost is the difference between a profitable stream and a budget overrun. Sony is among the enterprises delivering through the platform. ## How to scale WebRTC streaming to millions of viewers Three rules govern scale-out in 2026: - **Shard the SFU tree before you hit the wall.** Cascade SFUs across regions so each instance carries a bounded subscriber count. Treat the per-instance ceiling as a hard SLO input, not a discovery you make in production. - **Demote passive viewers off WebRTC.** Anyone who will not speak in the next 30 seconds does not need a peer connection. Move them to the HTTP path and reclaim the SFU capacity. - **Instrument the fallback rate.** A rising TURN-relay percentage is your early warning that a network region is degrading. It correlates with cost spikes before it shows up as viewer complaints. ## Failure modes worth designing for This is the section most comparisons skip. Real-time pipelines fail in specific, recurring ways. - **ICE restart storms.** A regional network blip triggers thousands of simultaneous ICE restarts, each renegotiating DTLS. Stagger reconnection with jitter or you DDoS your own SFU. - **Bandwidth estimator oscillation.** Aggressive probing on lossy mobile links causes the encoder to flap between layers, producing visible quality pumping. Cap the rate of layer switches. - **Repackaging buffer drift.** If the WebRTC-to-CMAF bridge under-buffers, viewers on the HTTP leg see stalls; over-buffer and you erase the low-latency advantage. Tune to your measured 95th-percentile jitter. - **TURN exhaustion.** A misconfigured firewall on the corporate side forces 100% relay for an enterprise audience, blowing your egress model. Monitor relay ratio per customer, not just globally. ## What to measure: 2026 threshold values - **Join time:** under 2 seconds, including ICE and DTLS handshake. - **End-to-end delay:** under 500ms for interactive WebRTC; under 3s for the LL-HLS leg. - **Packet loss tolerance:** usable below 1%, concealment-dependent up to ~5%. - **Reconnection time:** sub-second with ICE restart and pre-warmed candidates. - **Freeze ratio:** under 0.5% of playback time at the receiver. ## FAQ ### Is WebRTC streaming better than HLS for low latency? For true interactivity under 500ms, yes โ€” WebRTC delivers 50โ€“200ms glass-to-glass versus 2โ€“5s for well-tuned LL-HLS in 2026. But HLS scales to millions on stateless HTTP infrastructure, while WebRTC requires stateful SFU fan-out. Choose by whether viewers need to respond in real time or merely watch with low latency. ### What is the best WebRTC CDN for real-time video streaming? The right fit depends on your egress path. For WebRTC ingest with HTTP fan-out, a cost-efficient CDN handling the LL-HLS leg at high concurrency matters most. Evaluate per-GB cost at your peak volume, scaling behavior under spikes, and origin-shield efficiency rather than raw feature checklists. ### How many viewers can a single SFU handle? A well-provisioned SFU instance forwards to low thousands of subscribers per published stream before egress bandwidth and connection-state overhead force sharding. Beyond that, cascade SFUs regionally and demote passive viewers to an HTTP delivery path. ### Why does WebRTC use UDP instead of TCP? Real-time media values timeliness over guaranteed delivery. SRTP over UDP lets late packets be dropped and concealed rather than retransmitted, avoiding TCP's head-of-line blocking under loss. A 40ms-late frame is worthless; dropping it preserves the interactive feel. ### How much TURN relay traffic should I budget for? As of 2026, plan for 15โ€“25% of sessions to fall back to TURN relay due to symmetric NAT or restrictive firewalls. Relayed media is your most expensive byte, so monitor relay ratio per audience segment and treat sudden increases as a network-degradation signal. ## Your move this week Instrument the freeze ratio and TURN-relay percentage on one live session, measured at the receiver. If your relay rate climbs past 25% or your freeze ratio crosses 0.5%, you have a network or fan-out problem hiding behind acceptable averages. Run a synthetic scale test that demotes passive viewers to your HTTP leg and watch what it reclaims on the SFU tier โ€” then tell us where your per-instance ceiling actually landed. That number is the single most useful input to any real-time streaming capacity plan. Share: [f](https://www.facebook.com/sharer/sharer.php?u=https://blog.blazingcdn.com/en-us/cdn-and-webrtc-the-future-of-real-time-streaming) [in](https://www.linkedin.com/sharing/share-offsite/?url=https://blog.blazingcdn.com/en-us/cdn-and-webrtc-the-future-of-real-time-streaming) [๐•](https://twitter.com/intent/tweet?url=https://blog.blazingcdn.com/en-us/cdn-and-webrtc-the-future-of-real-time-streaming&text=) [โœ‰](mailto:?subject=%3Cspan%20id="hs_cos_wrapper_name"%20class="hs_cos_wrapper%20hs_cos_wrapper_meta_field%20hs_cos_wrapper_type_text"%20style=""%20data-hs-cos-general-type="meta_field"%20data-hs-cos-type="text"%20%3EWebRTC%20vs%20CDN%20Streaming:%20When%20Sub-Second%20Latency%20Is%20Worth%20It%3C/span%3E&body=https://blog.blazingcdn.com/en-us/cdn-and-webrtc-the-future-of-real-time-streaming) ![BlazingCDN](https://blog.blazingcdn.com/hs-fs/hubfs/Logo/blog-logo-w.png?height=24&name=blog-logo-w.png) *Heavy traffic.* Light bill. The CDN for video and large traffic Their monthly bill vs ours - 20 TBFastly $2,087**$92.50** - 50 TBCDN77 $990**$215** - 200 TBCloudFront $11,965**$765** Published list prices, Aug 2026 [Calculate your cost](https://blazingcdn.com/cdn-cost-calculator/?utm_source=blog&utm_medium=sidebar&utm_campaign=blog_sidebar&utm_content=compare_calc) ## Related posts [![](https://blog.blazingcdn.com/hubfs/Gemini-Blog/image-Sep-21-2026-07-30-27-5571-AM.jpeg)](https://blog.blazingcdn.com/en-us/tls-1-3-and-0-rtt-at-the-edge-the-real-handshake-cost?hsLang=en-us) Learn ### [TLS 1.3 and 0-RTT at the Edge: The Real Handshake Cost](https://blog.blazingcdn.com/en-us/tls-1-3-and-0-rtt-at-the-edge-the-real-handshake-cost?hsLang=en-us) TLS 1.3 removes exactly one round trip from a full handshake compared with TLS 1.2, and 0-RTT removes one more on ... Sep 21, 2026, 9:33:42โ€ฏAM [Read more](https://blog.blazingcdn.com/en-us/tls-1-3-and-0-rtt-at-the-edge-the-real-handshake-cost?hsLang=en-us) [![](https://blog.blazingcdn.com/hubfs/Gemini-Blog/image-Sep-20-2026-07-30-23-3710-AM.jpeg)](https://blog.blazingcdn.com/en-us/anycast-vs-dns-routing-how-a-cdn-picks-the-pop?hsLang=en-us) Learn ### [Anycast vs DNS Routing: How a CDN Picks the PoP](https://blog.blazingcdn.com/en-us/anycast-vs-dns-routing-how-a-cdn-picks-the-pop?hsLang=en-us) Evaluated February 2026. Two mechanisms decide which edge serves a request, and they fail on completely different ... Sep 20, 2026, 9:33:54โ€ฏAM [Read more](https://blog.blazingcdn.com/en-us/anycast-vs-dns-routing-how-a-cdn-picks-the-pop?hsLang=en-us) [![](https://blog.blazingcdn.com/hubfs/Gemini-Blog/image-Sep-20-2026-07-00-33-3709-AM.jpeg)](https://blog.blazingcdn.com/en-us/understanding-cloudflares-rate-limiting-pricing?hsLang=en-us) Security ### [Cloudflare Rate Limiting Pricing 2026: Plans, Rules and Real Costs](https://blog.blazingcdn.com/en-us/understanding-cloudflares-rate-limiting-pricing?hsLang=en-us) Cloudflare Rate Limiting Pricing 2026: Plans, Rules, Real Costs Cloudflare rate limiting pricing has one detail that ... Sep 20, 2026, 9:02:12โ€ฏAM [Read more](https://blog.blazingcdn.com/en-us/understanding-cloudflares-rate-limiting-pricing?hsLang=en-us) [![BlazingCDN](https://blog.blazingcdn.com/hubfs/Logo/blog-logo-w.png)](https://blog.blazingcdn.com?hsLang=en-us) [๐Ÿ“˜ Learn](https://blog.blazingcdn.com/cdn-learn?hsLang=en-us) [๐Ÿ“Š Benchmarks](https://blog.blazingcdn.com/cdn-ratings-and-benchmarks?hsLang=en-us) [๐ŸŽฌ Video & Streaming](https://blog.blazingcdn.com/video-streaming-cdn?hsLang=en-us) [๐Ÿญ Industries](https://blog.blazingcdn.com/cdn-industry-insights?hsLang=en-us) [๐Ÿ’ฐ Pricing & Costs](https://blog.blazingcdn.com/cdn-pricing-and-cdn-costs?hsLang=en-us) [โšก Compare](https://blog.blazingcdn.com/cdn-comparison?hsLang=en-us) [๐Ÿ”’ Security](https://blog.blazingcdn.com/cdn-security?hsLang=en-us) [๐Ÿ› ๏ธ Tools](https://blog.blazingcdn.com/cdn-tools?hsLang=en-us) [โœ๏ธ Publish with us](https://blog.blazingcdn.com/publish-with-us?hsLang=en-us) in f ๐• โœ‰ Copyright ยฉ BlazingCDN |. All rights reserved. ![](https://matomo.blazingcdn.com/matomo.php?idsite=1&rec=1) ```json { "@context" : "https://schema.org", "@id" : "https://blazingcdn.com/#organization", "@type" : "Organization", "description" : "High-volume CDN for video, live streaming, OTT/IPTV, software, games, SaaS and large-file delivery.", "logo" : { "@id" : "https://blazingcdn.com/#logo", "@type" : "ImageObject", "height" : 560, "url" : "https://blazingcdn.com/wp-content/uploads/2024/08/logo-560-560.png", "width" : 560 }, "name" : "BlazingCDN", "sameAs" : [ "https://www.linkedin.com/company/68261278", "https://x.com/BlazingCdn", "https://twitter.com/BlazingCdn", "https://www.facebook.com/BlazingCDN", "https://www.reddit.com/r/BlazingCDN/" ], "url" : "https://blazingcdn.com/" } ``` ```json { "@context" : "https://schema.org", "@id" : "https://blog.blazingcdn.com/#website", "@type" : "WebSite", "inLanguage" : "en-US", "name" : "BlazingCDN Blog", "publisher" : { "@id" : "https://blazingcdn.com/#organization" }, "url" : "https://blog.blazingcdn.com/" } ``` ```json { "@context" : "https://schema.org", "@id" : "https://blog.blazingcdn.com/en-us/cdn-and-webrtc-the-future-of-real-time-streaming#article", "@type" : "BlogPosting", "author" : { "@type" : "Person", "name" : "BlazingCDN" }, "dateModified" : "2026-09-17T17:10:24Z", "datePublished" : "2026-09-17T17:10:24Z", "description" : "WebRTC vs CDN-based HLS/DASH streaming compared on latency, scale, cost and reliability, with hybrid patterns for interactive live video.", "headline" : "WebRTC vs CDN Streaming: When Sub-Second Latency Is Worth It", "image" : "https://143144902.fs1.hubspotusercontent-eu1.net/hubfs/143144902/Gemini-Blog/image-May-29-2026-08-41-15-5646-AM.png", "inLanguage" : "en-us", "isPartOf" : { "@id" : "https://blog.blazingcdn.com/#website" }, "mainEntityOfPage" : { "@id" : "https://blog.blazingcdn.com/en-us/cdn-and-webrtc-the-future-of-real-time-streaming", "@type" : "WebPage" }, "publisher" : { "@id" : "https://blazingcdn.com/#organization" }, "wordCount" : 1648 } ``` ```json { "@context" : "https://schema.org", "@id" : "https://blog.blazingcdn.com/en-us/cdn-and-webrtc-the-future-of-real-time-streaming#breadcrumb", "@type" : "BreadcrumbList", "itemListElement" : [ { "@type" : "ListItem", "item" : "https://blog.blazingcdn.com/en-us", "name" : "Blog", "position" : 1 }, { "@type" : "ListItem", "item" : "https://blog.blazingcdn.com/en-us/cdn-and-webrtc-the-future-of-real-time-streaming", "name" : "WebRTC vs CDN Streaming: When Sub-Second Latency Is Worth It", "position" : 2 } ] } ``` ```json { "@context" : "https://schema.org", "@id" : "https://blog.blazingcdn.com/en-us/cdn-and-webrtc-the-future-of-real-time-streaming#faq", "@type" : "FAQPage", "mainEntity" : [ { "@type" : "Question", "acceptedAnswer" : { "@type" : "Answer", "text" : "For true interactivity under 500ms, yes โ€” WebRTC delivers 50โ€“200ms glass-to-glass versus 2โ€“5s for well-tuned LL-HLS in 2026. But HLS scales to millions on stateless HTTP infrastructure, while WebRTC requires stateful SFU fan-out. Choose by whether viewers need to respond in real time or merely watch with low latency." }, "name" : "Is WebRTC streaming better than HLS for low latency?" }, { "@type" : "Question", "acceptedAnswer" : { "@type" : "Answer", "text" : "The right fit depends on your egress path. For WebRTC ingest with HTTP fan-out, a cost-efficient CDN handling the LL-HLS leg at high concurrency matters most. Evaluate per-GB cost at your peak volume, scaling behavior under spikes, and origin-shield efficiency rather than raw feature checklists." }, "name" : "What is the best WebRTC CDN for real-time video streaming?" }, { "@type" : "Question", "acceptedAnswer" : { "@type" : "Answer", "text" : "A well-provisioned SFU instance forwards to low thousands of subscribers per published stream before egress bandwidth and connection-state overhead force sharding. Beyond that, cascade SFUs regionally and demote passive viewers to an HTTP delivery path." }, "name" : "How many viewers can a single SFU handle?" }, { "@type" : "Question", "acceptedAnswer" : { "@type" : "Answer", "text" : "Real-time media values timeliness over guaranteed delivery. SRTP over UDP lets late packets be dropped and concealed rather than retransmitted, avoiding TCP's head-of-line blocking under loss. A 40ms-late frame is worthless; dropping it preserves the interactive feel." }, "name" : "Why does WebRTC use UDP instead of TCP?" }, { "@type" : "Question", "acceptedAnswer" : { "@type" : "Answer", "text" : "As of 2026, plan for 15โ€“25% of sessions to fall back to TURN relay due to symmetric NAT or restrictive firewalls. Relayed media is your most expensive byte, so monitor relay ratio per audience segment and treat sudden increases as a network-degradation signal." }, "name" : "How much TURN relay traffic should I budget for?" } ] } ```