This informational document describes a tree-based CDN architecture
(called TreeDN) addressing the scaling challenges of live streaming to
mass audiences. The document points out three reasons why traditional
multicast solutions did not receive widespread deployment and use.
Afterwards the document argues that combining Source-Specific
Multicast (SSM) [RFC4607] and Automatic Multicast Tunneling (AMT)
[RFC7450] addresses all these challenges.

Even though the document is fairly short, it occasionally feels a bit
repetitive (e.g., first paragraph in Section 5) and sometimes a bit
too much white paper marketing style. While the authors claim that
TreeDN solves the chicken and egg problem of multicast deployments, it
feels to me that this document was written in order to help solve the
chicken and egg problem of TreeDN deployment. ;-) There is probably
nothing wrong with this kind of informational documents marketing a
certain architectural idea. However, as someone with a background in
research, I would have loved to see some hard (quantitative) facts
demonstrating that TreeDN is indeed a good solution outperforming
others and not suffering from the general multicast challenges.	Some
descriptions of several successful deployments of the TreeDN idea
would make this document stronger.