CDN Alliance’s Low Latency Working Group Whitepaper Lays Groundwork for Standardization

A Milestone for Low Latency Streaming

The streaming industry has long grappled with the challenges of delivering low latency at scale. Whether it’s the race to beat social media spoilers of a live sports broadcast, the need to sync remote and on-location bidders at live auctions, or the need to maximize revenue opportunities for live casinos or sports betting platforms, for many streaming verticals the ability to stream content in real time is no longer a luxury—it’s a necessity. That’s why the CDN Alliance’s recent whitepaper, Low Latency Streaming – A First Step Towards Standardization, marks a significant step forward for the industry.

At Ceeblue, we’re committed to optimizing and scaling existing real-time and sub-second technologies like WebRTC and HESP and to developing next-generation real-time streaming technologies like WebRTS, and we’re thrilled to see the CDN Alliance leading the charge toward clarity, collaboration, and standardization. As co-chair of the Low Latency Working Group (LLWG), Ceeblue has been deeply involved in the creation of this whitepaper from start to finish. Our engineers contributed real-world latency comparison data to help ground the findings in actual performance, and we spent many hours working alongside other LLWG members to align on the document’s focus, structure, conclusions, and final language.

This is Ceeblue’s core domain. We are low-latency specialists. Watching larger and more established companies attach the “Ultra-Low Latency” label to decade-old pipelines delivering 7-second streams is more than frustrating—it’s damaging. It distorts expectations and undermines the very real effort required to achieve reliable sub-second delivery at scale. This whitepaper is a first step in holding vendors accountable for the terminology they use and ensuring that buyers can confidently navigate the industry with a clear understanding of what different latency tiers and technologies actually mean in production.

Why Standardizing Low Latency Matters

One of the central issues the whitepaper tackles is the rampant ambiguity surrounding the term “Low Latency.” Different vendors, content providers, and platforms often have vastly different definitions of what qualifies as “low,” “ultra-low,” or “real-time” latency. We’ve seen vendors use “ultra-low latency” to describe video workflows with 7 to 10 seconds of delay—a range that clearly doesn’t support interactive applications or sub-second use cases.

This inconsistency doesn’t just confuse developers. It blurs expectations for everyone involved in the decision-making process: procurement teams, platform architects, CTOs, and even end users. It muddies the waters. It’s like reserving a premium suite where “premium” just means there’s a bidet in the bathroom. That kind of label inflation cheapens the real progress that engineering teams across the industry have made in pushing latency boundaries downward.

The LLWG whitepaper takes a direct approach to fixing this. It categorizes low latency into four distinct tiers and outlines the technologies associated with each. This classification provides a shared lexicon that allows stakeholders to align on expectations and make informed choices about technologies and workflows.

In doing so, it helps level the playing field: vendors who make exaggerated claims can be held to account, and those who provide real performance gains can be recognized accordingly.

Glass-to-Glass vs. Delivery-Only: Getting the Metrics Right

The LLWG advocates for a comprehensive, end-to-end definition of latency—often referred to as “glass-to-glass”—spanning from the camera lens to the viewer’s screen. This holistic approach, which breaks down the video pipeline into first mile (camera to CDN), middle mile (transcoding and CDN distribution), and last mile (CDN to viewer), provides a much-needed framework for assessing true performance.

This stands in contrast to vendor-specific definitions that may only focus on delivery latency, creating unrealistic expectations or misleading benchmarks. One vendor may claim 2-second latency by omitting the ingest time or skipping over the player buffer, while another may advertise 3 seconds for the exact same end-to-end delay.

Even industry coverage has struggled to keep these definitions consistent. Articles from different publications have used “low latency” to describe both LL-HLS workflows at 8–10 seconds and WebRTC pipelines running under 500 milliseconds.

With this whitepaper, the LLWG sets the stage for a shared understanding that empowers developers, platforms, and content providers alike.

Four Flavors of Latency: LL, ULL, SSS, and RTS

A particularly impactful contribution of the LLWG whitepaper is the delineation of four key latency categories:

• Low Latency (3–12 seconds): Ideal for linear broadcast and non-critical live content. Often delivered over LL-HLS or LL-DASH.
  
• Ultra-Low Latency (1–3 seconds): Targets OTT sports and eSports broadcasts with slightly higher viewer expectations.
  
• Sub-Second Streaming (0.5–1 second): Enables interactivity and audience participation for events like trivia games, live concerts, or town halls.
  
• Real-Time Streaming (under 0.5 seconds): The gold standard for use cases like sports betting, auctions, and interactive live shows where milliseconds matter.
  

This structure gives everyone in the streaming supply chain a common vocabulary to work from. Whether you’re building a tech stack, evaluating vendors, or planning a product roadmap, this classification enables more informed and transparent decision-making.

Matching Tech to the Use Case

The whitepaper also does an excellent job mapping these latency categories to corresponding technologies. From legacy options like LL-HLS and LL-DASH, to sub-second and real-time protocols such as WebRTC, WebSockets, and Ceeblue’s own segmented WebRTS, the document offers a comprehensive breakdown of what’s possible—and what’s still emerging.

It also discusses the downsides of the legacy technologies, like LL-HLS, which doesn’t scale due to its inefficient abuse of the origin and which can’t achieve latencies low enough to support true interactivity. The best-known lowest-latency options also have their drawbacks, like WebRTC’s difficulties with corporate network security and complex scalability paths or HESP’s vendor lock-in. 

What Makes WebRTS Different

WebRTS stands out because it addresses many of the limitations of these legacy and sub-second or real-time protocols. It was designed specifically to deliver sub-500ms video, with no VOD baggage. It runs on standard CDNs, with optional support for DRM, adaptive bitrate (ABR), and cross-platform compatibility.

Where WebRTC often struggles to scale beyond a few thousand concurrent viewers, and where WebSockets require custom backends and lack robust CDN support, WebRTS provides a practical middle ground. It leverages standard web infrastructure and caching systems while maintaining real-time delivery goals.

WebRTS also benefits from being open-source. That means platforms can integrate and experiment without vendor lock-in, while also contributing back to the standard to push it forward.

Use Cases That Can’t Wait

The whitepaper illustrates just how varied—and critical—low latency use cases have become:

• Sports OTT broadcasters racing to avoid spoilers,
  
• In-stadium viewing needing split-second streaming,
  
• Betting platforms demanding real-time odds and interactions,
  
• Live shopping platforms delivering organic, personalized shopping experiences,
  
• Live auctions where remote and in-person bidders need to be synced.

The need for low latency streaming is as diverse as it is urgent

In the case of sports betting and auctions, the economic incentive is paramount. As is commonly heard in this industry: “If there’s an issue with the stream, no one gets paid.” The entire event hinges on the real-time availability and reliability of the broadcast. If there is a question about the integrity of the broadcast, bets must be voided and auction winners can’t be confirmed.

Perception vs. Reality: Lagging Expectations in the Industry

While some industries have embraced real-time streaming, others still treat it as a luxury, nice-to-have feature. OTT broadcasters often assume that 15–30 seconds latency is just unavoidable, and that anything beyond that is science fiction. But when we tell people in the industry that the Super Bowl is still broadcast with latencies approaching a full minute, they’re genuinely surprised.

The LLWG whitepaper helps shift this perception. By clearly laying out what technologies can achieve today, it helps set a higher bar across the industry.

Procurement, RFPs, and Smarter Technology Selection

By standardizing definitions and categories, the whitepaper helps streamline technology evaluation and procurement.

Platform teams evaluating proposals can now ask specific questions: Does this solution meet Sub-Second or Real-Time latency? How is latency measured? Can it support ABR or DRM?

Procurement teams can stop relying on marketing buzzwords and start assessing whether solutions actually meet the use case requirements. This increases transparency and helps buyers make faster, better-informed decisions.

Open Standards Promote Adoption

One of the reasons HLS and DASH saw such widespread adoption was their openness. They were royalty-free, widely supported, and benefited from large communities of developers.

WebRTS is built with that same philosophy. As an open-source framework, it offers transparency, flexibility, and the ability to evolve. Ceeblue is committed to maintaining this openness as WebRTS grows.

MoQ: Tremendous Promise, Uncertain Path

Media over QUIC (MoQ) is an emerging technology with real potential to reshape the streaming stack. It could eliminate the need for intermediate repackaging steps and offer native support for low-latency, reliable, scalable delivery over QUIC.

That said, the protocol is still under development. It carries the potential to become groundbreaking … or to be weighed down by “development by committee.”

The LLWG has chosen not to force-fit MoQ into existing categories prematurely. Instead, future versions of the whitepaper will revisit it as the protocol matures and real-world deployments become available.

The Path Forward

The CDN Alliance’s LLWG makes it clear: this is just the beginning. Standardizing terminology and clarifying technology capabilities are essential first steps, but the journey doesn’t end there. Ongoing collaboration from vendors, developers, CDNs, and content platforms is critical to refining this resource and keeping it up to date.

At Ceeblue, we believe that real-time video should be as seamless and accessible as text or static images, and that starts with a shared foundation. The LLWG whitepaper is that foundation. It arms the community with the vocabulary, frameworks, and insights needed to move the industry forward.

We encourage everyone in the streaming space, from developers to business leaders, to read the whitepaper, engage with the LLWG, and take part in building a future where anything that isn’t low latency is a novelty.

You can explore the whitepaper and learn how to get involved with the Low Latency Working Group at CDN Alliance’s official page and learn more about WebRTS using the links below.