##Organization
Reviewer 1
The organization of the manuscript should be improved.
Authors start by discussing in some detail key technologies and specific tools (LLTD, eSPID, SCBP…) too early (in the introduction), then discuss related work already on a very specialized perspective (bandwidth prediction for flows and PBAC/MBAC) and only afterwards provide their own contributions - also following a very specialized and modular organization.
In my opinion the problem statement is not clear. Authors should discuss, in the introduction, which are the goals of their proposal and why is the proposal is relevant.
As already mentioned, the state of the art is focused on a subset of the technologies/components included in their proposal, without analyzing previous approaches to the general problems they are addressing.
Section 3 (contributions) starts by discussing the authors’ contributions to specific components of their framework, instead of providing first an overview of the framework as a whole. Besides, at some points the content is excessively detailed without conveying relevant content to reader (I do not understand, for instance, the relevance or the need to present in such detail the QSLP-LAN message formats).
The second part of the evaluation is interesting to understand how the system works, but does not actually provide evaluation. For the same scenario would it be possible to achieve similar performance using alternative approaches? Does the system scale well with a higher number of nodes (and, more important, higher number of flows)?
Overall the paper seems fragmented, with detailed discussions of individual components of the proposed framework but without a proper description, discussion and evaluation of the framework as a whole.