3 Vendors Pin Hopes on 1-Way Real-Time Streaming over HTTP
Proprietary Solutions Seek Market Sweet Spot at Costs Below MOQ & WebRTC
By Fred Dawson
MOQ may be moving rapidly toward adoption as a real-time streaming standard with growing industry support, but that’s not deterring some suppliers from ramping up efforts to promote new approaches to streaming at sub-second latencies over existing infrastructure.
Ceeblue, nanocosmos and Nakadi are betting ready-to-deploy real-time streaming that works with software tweaks on conventional CDNs will outweigh people’s reluctance to go with proprietary solutions. Afterall, they argue, their approaches are far cheaper to implement than MOQ and WebRTC platforms that require new infrastructure.
“People don’t want to have to operate over different CDNs from the ones they’re using now or build new ones in the case of companies streaming content over their own CDNs,” says Jonas Blotz, vice president of engineering at Ceeblue. As previously reported, Netherlands-based Ceeblue is the developer of the ultralow-latency Web Real-Time Streaming (WebRTS) platform, which lately has begun drawing customers to the idea they can hit half-second latencies over existing HTTP infrastructure with the aid of a software upgrade to CNN hardware.
Their biggest announced win so far is with the wholesale arm of the French mobile giant Orange, which is offering the company’s retail service outlets the option to run their streaming services over WebRTS on the Orange Wholesale CDN. According to Josh Bergen, business operations coordinator for Ceeblue, WebRTS has also been integrated to run on CDNs supported by Varnish Software, and Akamai has made support for WebRTS available to one of its customers in conjunction with the company’s previously reported positioning as a supplier of a video-optimized global cloud computing platform that can be used to activate home-grown CDNs.
As we also reported earlier this year, nanocosmos, too, is promoting a new approach to real-time streaming, which is a Media-over-QUIC solution that’s not compatible with MOQ. nanocosmos CEO Oliver Lietz doesn’t hesitate to dispel rumors circulating at the recent IBC show that his company was preparing to ditch its approach to using an HTTP-compatible version of QUIC in favor of a shift to the MOQ platform.
“We’re not moving in the direction of compatibility with MOQ,” Lietz says. “People want to build business on a stable system. It’s not clear where MOQ is going.”
Just now entering the fray, the little-known Nakadi debuted at IBC as another player with a software-engineered approach compatible with traditional HTTP streaming infrastructure that supports real-time streaming. But even though, according to Nakadi founder and CEO George Mikeladze, Nakadi’s Fusion technology supports “latencies like WebRTC,” he believes MOQ will eventually prevail.
As a result, Nakadi sees Fusion as a precursor to MOQ, notwithstanding the incompatibilities. “We’re using TCP now,” Mikeladze says. But when the time is right, “we’ll be running on MOQ Transport.”
While these companies have convincingly demonstrated their platforms perform as claimed, they seem undaunted by the fact that they don’t support multidirectional streaming. Speaking with us about WebRTS at the NAB Show in April, Ceeblue COO Lawton Chene said the protocol could be made to work multidirectionally if needed, but he didn’t see such capabilities as a priority among sports broadcasters and others who are looking to stream in real time at mass-audience scales over existing CDNs.
For its part, nanocosmos is getting plenty of business with its approaches to one-way streaming, Lietz says. He notes the company’s new MoQ solution, which was specifically designed to cut latencies on streams before they are ingested onto the company’s global cloud CDN and after they exit onto last-mile networks, fits right in with a targeted market of enterprises looking for real-time streaming serving no more than 200,000 simultaneous users.
But Lietz does note that the new technology, which is also being implemented on the cloud backbone, can be used to expand the targeted real-time streaming market to users who want to reach much larger audiences. “We’re talking to BT, which is considering MoQ as a replacement to HLS,” Lietz says.
Customers on the nanocosmos cloud platform, which charges the same for use with or without its version of UDP-based MoQ, can use the new mode whenever they desire, which owing to the switch from TCP-based last-mile streaming beyond the CDN, might require resetting licenses. “There’s no forklifting if you want to switch to MoQ, which allows us to push innovation as first movers in real-time streaming,” Lietz says.
As for support for interactivity, nanocosmos specializes in enabling the simultaneous core content reception essential to interactive apps. But the focus is on facilitating the real-time data exchanges that service providers and other enterprises are looking for in apps like micro-betting, e-commerce and live online auctions.
Interactivity is far from being a priority at Nakadi. While Nakadi’s Fusion platform enables one-way streaming in real time, making that possible wasn’t the motivation behind its development nor is it a selling point in the messaging Nakadi generated with rollout at IBC.
Instead, Mikeladze says, the low-hanging fruit for Fusion is the market’s need for HTTP-based streaming that can deliver high-quality video at bandwidth low enough to reach billions of people who are limited to accessing the internet on cellphones that can’t stream video at decent quality or even at all. “We’re streaming at ten times the quality at 10X lower costs,” he says.
Nakadi is bringing Fusion to market as the heart of a fully managed white-label OTT service, which can be run in the public cloud, on premises or in hybrid mode. The company says Fusion is “a pre-standard Media over QUIC-based innovation designed to solve the persistent cost, quality, and scale limitations of WebRTC, HLS, DASH, and CMAF.”
At IBC, the company conducted a live demonstration of two side-by-side streams of content fed from New York City to Amsterdam, one via HLS the other via Fusion, to confirm its claim that streams running on the software can maintain much higher quality compared to basic HLS when both encounter the same level of network turbulence.
For example, with interference raised in the test to produce a 10% packet loss rate on both feeds, the Fusion stream, which automatically makes ABR adjustments when quality is diminished, continued delivering its high-quality 1080p HD feed at 20mbps with no disruptions whereas the 1080p stream operating at 5mbps over HLS was forced to cut back to 360p at 1mbps to avoid the disruptions occurring at 5mbps.
So far, Nakadi has not provided any documentation or public statements explaining mechanisms it uses beyond QUIC to achieve touted performance parameters. But it does make clear that it employs an API as “a generic and content-agnostic event broker” that encapsulates all aspects of stream processing in microservice modules that conform to schemas defined by user-selected “EventTypes.”
In the process, redundancies are eliminated as is superfluous information about “the backing messaging infrastructure.” This appears to be analogous to what Ceeblue does with reductions in player/server communications data related to changes in manifest files.
“We’re not giving away our secrets,” says Nakadi CTO Thierry Fautier. Fautier, a well-known former strategist for Harmonic, has taken on the role with Nakadi while remaining affiliated with The Media League, a consultancy he co-founded in late 2024.
Nakadi, while new to the marketplace under current branding, is related in some way to Qarva, a company founded by Mikeladze that made its mark in the Eastern European Republic of Georgia providing IP streaming support for a once-floundering pay TV service. It’s unclear what the relationship between the two companies is, what investment backing Nakadi has or whether Qarva, which has a website dated 2022 and last made appearances at IBC, NAB and other trade shows in 2023, is still operating.
Nakadi’s initial focus is on broadcasters, operators, and media companies in emerging regions, where the economics of infrastructure matter most. However, Mikeladze says the solution is equally suited for mature markets, where streaming services are commoditizing and cost efficiency has become the next competitive frontier. Along with operating at ultralow latencies, the company says the platform supports persistent quality streaming at resolutions as high as 4K and 8K with scalability to millions of concurrent users.
The long-term successes of these companies’ real-time streaming solutions are likely to hinge on three things: overcoming market reluctance to rely on proprietary solutions, how an inevitably diminishing cost differential will impact their standings as MOQ takes hold at massive scales, and how important support for full-scale multidirectional streaming becomes as the socialization of video usage in all domains gains traction.
The good news for users is the options now at hand to put real-time streaming to use go far beyond what was the case just a little over a year ago when the choices largely revolved around WebRTC-based platforms. But given that this expanded supplier ecosystem now also includes MOQ-based solutions, which, as previously reported, are being put into play for pre-standard experimentation on several fronts, it remains to be seen how good the news is for the providers of proprietary one-way real-time streaming solutions.
Jonas Blotz, VP, Engineering, Ceeblue