A Post-IBC MPEG-DASH Status Check

Article Featured Image

While most of the buzz at IBC in Amsterdam last month was generated by the shiny HEVC + 4K couple, it was a good occasion to step back from the hype and measure how one of the most important video standards (at least for the OTT world) had spread over the industry since last year. MPEG-DASH is not sexy (it’s basically just a collection of XML schemes), it’s a bit boring (as it’s usually demonstrated with the "Big Buck Bunny" clip that we've all seen no fewer than 2.000 times), and it’s complex. But once you go past all that, you might find that its potential to simplify your workflows and boost the profitability of your video service is high.

And that’s precisely what the industry has finally understood: Video distributors want to streamline their workflows and save their storage budgets by reducing the number of output formats, so they search for ways to build long-term strategies with evolutionary solutions—all of which implies moving away from proprietary ABR solutions like Smooth Streaming, HLS, HDS, and the now defunct Widevine WVM proprietary packaging format (rest in peace), DASH’s first victim.

Contrasting Situations

Two and a half years after DASH became a Draft International Standard, the industry is balanced between hope and skepticism. On the positive side, we see that DASH, being supported in roughly a hundred solutions, has really made its way outside the labs into the real production workflows and gathered a wide industry consensus with strong supporters like Microsoft, Netflix, and Google. The growing number of DASH Industry Forum participants is a good indicator of traction, and we see some deciding factors shaping up for boosting DASH uptake: HbbTV 1.5 is on the verge of becoming a reality in Europe and DASH as an HEVC transport layer is catching on in Japan. Widevine’s quiet but steady move of dropping their proprietary packaging for DASH is a really strong signal for the rest of the industry, as it is showing that today there is no strong market position that can justify getting bogged down in futureless architecture choices (although some might still not have gotten the memo).

On the dark side of the force, the DASH supporters are facing a not-so-promising field situation where deployments are still prospective and the 30-days-trial a reality: People are still evaluating the technology and making up their mind on which profile to use, heating up the calculators to check how they can really save on their current budget given the reality of the DASH+CENC+Multi-DRM support on the client side today. Let’s face it: Migrating to DASH for existing services might not be really compelling because the cost of this migration, and the difficulties of achieving success on all target devices makes it a no-brainer. Existing services would rather continue with their  HLS/HDS/Smooth Streaming workflows, which are stable and roughly interoperable inside their own subdomains; that’s a kind of reassuring "Yalta equilibrium" for many of us.

The question is different for HbbTV services. They don’t have any other choice but to implement DASH now. Early HEVC adopters are in the same boat, as using HEVC in ISO BMFF profile is the natural way as of now, though we might see alternative transport mechanisms finally challenging DASH in the upcoming years (see for example Haivision SRT). To summarize, H.264-based HbbTV and HEVC are currently the main drivers for DASH adoption, but this is not sufficient to accelerate the deployment rate right now.

What’s Missing?

For the premium OTT services, the most important penalty factor might be that Common Encryption + Multi-DRM is still very emergent (more on that later) and therefore they cannot leverage it yet to mutualize their packaged output between all their devices, which generally require 2 to 3 different DRMs for full coverage. Once CENC+Multi-DRM is widely supported on both content preparation and client-side ends, there is no doubt that the boost will be massive as the budget savings on storage and the gain in workflow simplicity will be HUGE. Two other important interop bricks are missing in DASH today: key exchange mechanisms between DASH packagers and DRM servers, a painful track on which the DASH-IF is currently working, and a standardized ingest protocol. I questioned encoder/transcoder/packager providers about their preference on this and usually they mention WebDAV first. The only dissonant voice came from the Unified Streaming team, who said that the Smooth Streaming ingest protocol was the most efficient and widely used in the industry and therefore could easily be patched for DASH ingest. My take on this is that the current situation will persist, meaning that big CDNs will decide which ingest protocol/encoder authentication methods they prefer, and the rest of the industry will comply. Hopefully no new protocol will be invented on the way…

Read the full article on blog.eltrovemo.com

Streaming Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues
Related Articles

Dynamic Adaptive Streaming over HTTP (DASH): Past, Present, and Future

As DASH begins to find adoption, it's useful to take a look back at the development of the standard, as well as to try and determine the issues that will shape the future