IBC Hackfest 2015 - Amsterdam

Vidlytics

 

 

 

 

Vidlytics is an attempt at an answer to the question of end-viewer quality of experience when streaming videos. We have attacked this through a three-pronged approach:

 

QoE Triangulation

Actively discovering viewer connection problems by comparisons between multiple route mid-points.
Manifest Engineering

Strategicially injecting, proxying and redirecting to gather media request analytics with minimal client disruption.
Analytics

Real-time visual representation of a user's video experience by leveraging data agregated from our servers and the client's themselves.

 

Use cases:

  • CDNs,
  • Content providers,
  • ISPs

 

Inspired by the Streamzilla challenge

 

Images at:  http://imgur.com/a/7VLRo

Github