Live Streaming Technic Comparison
Live Streaming, HLS, RTMP, and WebRTC
This tech comparison will guide you to the best choice between live-streaming solutions:
To build live-streaming system for your business, the right technic choice is key to success.
For a typical business model with thousands of active users, there are plenty of open-source solutions to choose from.
You may heard of
WebRTC. They are all for live-streaming, but what’s the difference? Let’s have a look at each other.
- Latency : Medium. Typical latency is 2~5 seconds, we can reduce the latency to around 1s.
- Popularity : Most of the apps in the store use
RTMPnowadays, big names include
Facebook Live, etc.
- System compatibility : There are many open source libraries and SDK to integrate with the app. But for the website, the client will have to support
Flash— browsers like mobile
Safaridon’t support it.
- Cost to scale : Medium.
- Latency : High. They are both chunk based format, which means new segment is ready only after a fixed interval. Typical latency is between 5~60 seconds.
- Popularity : Widely used in most of the streaming platform, websites and apps. TVs use
HLSbecause they don’t need real-realtime-live-streaming.
- System compatibility : Browser-friendly, easy to embed into webpages and apps,
Safarihas native support with
- Cost to scale : Low. They are transport over http, and easy to scale with
CDNs, which will reduce the total cost.
- Latency : Low.
- Popularity :
WebRTCis used in live chat and online meeting.
- System compatibility :
WebRTCis currently supported by
Mozilla Firefox, and
Microsoft’s Internet Explorerand Apple’s
Safarihave yet to add support for
WebRTC, support for these browsers comes in the form of 3rd party plugins right now.
- Cost to scale : Hard to scale. Some solutions will use
WebRTCfor input and
RTMPis a good choice which balance performance and cost.
DASHare cost efficient solutions support scale to large numbers of clients, with the disadvantage of notable latency.
WebRTCif for real live-streaming, but is hard to scale (at 2018).