In our blog series Meet the Minds Inventing the Future of Video, we speak with some of Ateme’s brightest minds to find out what they’re working on.
As part two of the series, Director Technology and Standards, Research and Innovation, Christophe Burdinat talks about the meaning, the challenges, and achievements of mABR.
Update: Congratulations to Christophe Burdinat, who has recently resumed his role as Chairman of the MABR task force of DVB’s CM-I group!
What is your role at Ateme?
A year-and-a-half ago, I joined Ateme as Director Technology and Standards, within the Research and Innovation team, headed by our CTO.
My role is split between two activities. The first is standardization — scouting and contributing, collaborative research project management, and experimental development. Secondly, I help disseminate our technologies and findings in papers, webinars, etc. I appreciate the transversal nature of this role. This allows us to help shape our industry through standardization organizations on one hand, and delve into advanced technologies on the other. In total, I focus mainly on video distribution aspects including CDN, multicast ABR, and integration in 5G networks.
What in particular have you been working on at Ateme?
For some time now, I’ve been working on the support of multicast/broadcast services within mobile networks. My arrival at Ateme coincided with increased distribution research. So, my first project here was to link a 5G setup to showcase the delivery of live content over 5G broadcast, leveraging the ROUTE protocol from ATSC 3.0.
When it comes to 5G, I’m a regular delegate of 3GPP. I also attend the 5G Media Action Group (5G-MAG) — a key industry forum where open-source reference tools are developed for the 5G media streaming architecture. Also, I am part of the Open Caching group of the Streaming Video Technology Alliance. We adapted the Open Caching API for Ateme’s NEA CDN solution. Moreover, we continue to ensure the solution is compliant with the Open Caching standard as the specifications evolve. Last but not least, I oversee the standardization of multicast ABR (mABR) at DVB.
What is mABR?
MABR, or Multicast Adaptive Bitrate, is an architecture and a protocol stack to deliver ABR video services (DASH, HLS, etc.) over multicast. This means delivering a single stream to multiple targeted viewers at the same time. This is a key technology for the convergence of broadband and broadcast. Several applications can be considered, depending on the network and the deployment model. Hence, for managed networks, it can be viewed as the next-gen IPTV. This is where operators shifting from MPEG-2 TS would be able to use the same headend for both live and on-demand/catch-up TV. It’s a way to scale a CDN for live content, for which no additional edge caching capacities would be required to absorb peak audiences. In addition, an mABR stack is also relevant for any IP-centric broadcast network, whether it be DTT (ATSC 3.0), satellite (DVB-NIP), mobile (5G Broadcast), etc.
What are the challenges Ateme has helped solve when it comes to mABR?
One of the main challenges is the delivery of low-latency live services over multicast. ABR services are served as sequences of media segments, available as small files containing a few seconds of content. Therefore, to offer latency comparable to traditional broadcast, the delivery of these media segments needs to start while they are still being produced. This leads to strict requirements on the full delivery chain, from headend ingestion to multicast reception in the mABR gateway. This is handled in ATSC 3.0 with the ROUTE protocol that Ateme is experimenting with in its own implementation, as well as with the GPAC open-source toolkit.
When mABR is used in a broadcast context, several services may compete for bandwidth in a fixed slice of spectrum. In this case, the concept of statmux is still as relevant. To achieve both low latency and statmux support, it is necessary to tightly integrate the mABR server function with the encoder and the packager. Because Ateme controls the end-to-end technology stack, we are uniquely positioned to ensure this tight integration.
What have you achieved in this field at Ateme?
At Ateme, I’ve been working on a 5G broadcast setup to experiment with and showcase different mABR stacks. Working on IP-centric networks, I haven’t seen any major difficulties when switching from one stack to another. Working with several stacks helps when addressing multiple markets. For instance, TV3.0, the Brazilian standard, has adopted the ROUTE protocol from ATSC 3.0. The DVB-MABR standard itself offers 2 profiles — ROUTE and FLUTE, from 3GPP. First-market implementations are based on FLUTE. MABR is included in several collaborative research projects. For example, it has been recently achieved in 5G Vista, which offers in-stadium live multi-camera views over 5G Broadcast. It is also used in Nested, which targets an energy-efficient solution for streaming over 5G. Leveraging mABR is a part of the Ateme approach to making video delivery more sustainable.
What does this change for viewers?
MABR can provide an enriched live-tv experience. This is accomplished by serving hybrid broadband/broadcast services, where additional components are made available over unicast. For example, in the case of audio/subtitle tracks for less popular languages. Support of rich hybrid broadband/broadcast services is at the core of a new study at DVB: Object-Based Media. Associated components could be metadata, commentary tracks, other media objects (trailers/making of/reviews), social media links, etc.
In an IPTV context, mABR would offer a high level of Quality of Experience to all devices at home, not only the TV. When deployed over mobile broadcast/multicast, the service can be offered to a large audience without any congestion. For instance, think about the stadium use case. In conclusion, MABR is a logical step indeed in video distribution, where streamlining and sustainability go hand in hand. This provides viewers with an enriched experience and better quality viewing.