Panel Discussion Low Latency, Ultra-Low Latency, And You
Sponsored by nanocosmos.
What technical and business factors should you consider in order to get your latency where you need it to be? And what strategic decisions and technology choices can you make to get there? There’s no one-size-fits-all answer to that question; it all depends on your content and your audience.
Jason Thibeault, Executive Director, Streaming Video Alliance, USA
KEY TAKE AWAYS
- What latency does my use case require? Latency always depends on the use case. There are use cases where it is nice to have ultra-low latency live streaming and there are use cases where it is a must. For interactive use cases e.g. live auctions, live betting ultra-low latency is a must. Hence, latency is not something that every streaming video use case can benefit from. It is very important to understand the use case, business model and user experience and level of interaction required.
- New use cases have emerged e.g. watch parties to engage fans remotely during a concert with a Q&A and the noticeable increased focus on player interactivity for specific verticals, but also for generic webcasts including interaction tools like polling and live chat, especially around live sports, betting, auctions and town hall meetings.
- The future of live streaming: Metrics offer additional information, intelligence to understand and better control the live streaming workflow. A global scalable CDN including the right tools improves insight on the quality of service and quality of experience. Metrics and Analytics are crucial for a QoS and QoE, like buffering, fps, video quality, successful viewer experience.
- Latency streaming & buffering: Intelligence on the player and the server part is required to optimize this and the stream needs to be of the highest quality as possible. It is important to understand the use case and help the customers to optimize. Trade offs need to be carefully evaluated.
- Is 5G a game changer for interactive use cases? For high quality you need higher bandwidth, so 5G is important. On the distribution side 5G might not always be sufficient if you are on the go for instance. On the other hand user generated content can benefit from it.
- HL5 & protocols: You need to consider scale, compatibility devices, plus there is an increased demand for browser-based use with instant access, on any device. Flexibility depending on the set up is required. For the architectural considerations (such as scale) it is not only the protocol, but the whole application that we need to look at. It should work on all devices, preferable directly in the browser, also in hostile networks, firewalls, wifi, mobile, etc.
- Hybrid delivery architecture: commercial & peer to peer: Peer to peer environments will always add to the latency. The important question to answer is if it is worth the cost in terms of higher latency.
- HLS / what format, what packaging is best? There is no one size fits all answer. Streaming optimization is a comprehensive challenge, all factors have to stay in balance. All optimization changes can have an impact on latency. If you look at one change, you always have to look at the impact on all other factors that your change might have. The whole workflow is important, starting from the camera, encoder, first mile, middle mile, last mile, delivery, playback. The devil is in the details, so full control and insight is required, incl. metrics
- “Reduce latency” How? Own as many parts of the workflow chain as you can. The more you can control, the better you can optimize. Also: understand the use case first, then determine latency. A clear understanding of the monetized business model is required for a successful business roll-out.
Speaker: Oliver Lietz, CEO, nanocosmos.
Oliver Lietz founded nanocosmos in 1998 with a strong background in video R&D for video/audio coding and streaming for the broadcast industry. Since then he has been focusing on developing solutions that integrate video technology into custom business applications like iGaming platforms.