##Novelty
Reviewer 1
This paper contribution is incremental, and concepts presented build on earlier QoS work such as topology discovery, flow identification/classification, flow bandwidth prediction and policing. By taking on too many of these fundamental aspects of traffic engineering, the paper struggles to keep a balance of depth in any specific topic. The novelty of the work is not clear, although a complete system consideration is presented.
The proposed schemes of ‘cooperation’ of all hosts for reservation of QoS needs for applications has been proposed previously, and has been a challenge to implement in the Internet. The same is the case for MBAC schemes that are prescribed again by the authors. The authors do not have clear justifications why this time around, their scheme would be implementable, even within LAN where applications need to access Internet to obtain services and content for their applications such as IPTV and HD video streaming.
More convincing novelty and a clear focus on the paper contribution could have made the paper more appealing.
Reviewer 2
While the solution itself it not novel, its application to the home environment (and unmanaged LANs in general) has some novelty.
QoS/QoE management in home networks is a relevant issue that affects telecom’s business (providing IP-based TV, VoD and voice services to their customers) and over-the-top services. There are several industry and academic proposals for QoS management in these scenarios, mostly based on DiffServ. This paper proposes a somehow different approach, but in my opinion it is still not convincing.