This is the final blog in our series that has focused on the launch of our new innovation StreamPilot. Over the four posts we’ve looked at the challenges in multi CDN environments associated with how sessions are set up for initial CDN selection, how quality of experience (QoE) can be measured server side for OTT streaming, the previous ways of doing in-session CDN selection, and finally, how to do in-session CDN selection.
StreamPilot has been specially engineered to overcome these challenges. Let’s look at how it does this!
As covered in earlier blogs, the current methods to manage a multi CDN delivery have severe limitations. The decisions are often based on old data coming from a subset of clients and the selection methods are slow, lack granularity for both measurement and selection or involve building complex proprietary clients. To make decisions in real time requires a fine granular visibility into all clients’ experience and the means to, without delay, move streams to the optimal delivery source. The lack of control and visibility results in slow and imprecise management of the delivery mechanism. This, in turn, makes it impossible to quickly identify and fix delivery problems that cause a bad quality of experience (QoE) for the end user.
A wish list of key requirements for the ultimate CDN selection functionality would include:
Edgeware’s StreamPilot is a cloud-based platform for multi-CDN management using a new method we call ‘segment redirect’. Using the HTTP-redirection mechanism for each segment while having the session tied to the StreamPilot service for its full duration puts StreamPilot in a unique position to both supervise all requests in real time and have the means to individually redirect each request.
Similar to software defined networking (SDN), where the control plane is separated from the data plane, StreamPilot uses segment redirect to insert itself into the control plane between the CDN and client without taking part in the data path. Here is how StreamPilot works.
All methods come with pros and cons. A ping time will be added to all segment requests, and there will be some added bandwidth when the requests are sent through StreamPilot, but that is typically 0.1% the streamed data.
Client-side monitoring can give more details, especially about user interaction and video playback, but there is plenty of information that can be extracted and interpreted on the server side. Furthermore, server-side monitoring works for all clients without any modification, and it is on the server side that the information should be gathered for decisions such as CDN selection.
Introducing a delivery and client independent service in the control plane that supervises and manages the delivery, is a powerful and flexible method that, after weighing up the pros and cons, comes out as a great strategy for managing a TV service. Without dependencies on any part of the delivery chain, it is a powerful platform to add new services to.
Kalle Henriksson, Founder, Edgeware
Torbjörn Einarsson, Expert, Media protocols and codecs, Edgeware
Do you want to know more? Fill out the form below and we will get in touch with you.