Creating WebRTC applications is simple, but making them scalable in a robust and efficient way isn’t. Due to this, many application developers do not dare to use their own infrastructures and instead prefer to leverage third party clouds for creating their scalable WebRTC services. These clouds are private and usually expose some kind of API that developers consume for accessing WebRTC communication capabilities such as group-communications, media recording or media transcoding. This model is very convenient for creating simple vertical applications, but it also has some drawbacks. First, these clouds billing models are based on pay-per-minute schemes that are not always compatible with the business models desired by application developers. Second, the exposed APIs tend to be restrictive and limit the freedom of developers for innovating. Third, these clouds, being fully private, cannot be extended or customized for specific needs developers may have.
In this talk, we propose an alternative solution to private WebRTC PaaS models based on leveraging novel cloud orchestration technologies such as Cloud Formation or Heat Templates for enabling developers to create their very own platform on top of public clouds. We analyze the technological ingredients required for enabling such models and present some experiments and production deployments showing how any developer may bring her own IaaS cloud for growing on it a full featured WebRTC platform with elastic scalability and full control on the billing and on the underlying technology evolution.