Distributed Bandwidth Allocation Strategy for QoE Fairness of Multiple Video Streams in Bottleneck Links
Round 1
Reviewer 1 Report
In this manuscript, the authors designed a fair bandwidth allocation based on congestion control to maximize the QoE fairness among video users in such a way as to maintain the overall QoE efficiency and network bandwidth utilization of the video system at an acceptable level.
However, the following concern should be addressed:
1- The related work should cover recent studies in the addressed area. The majority of papers are old.
2- It is strongly recommended to add a motivating example to illustrate the impact of the proposed approach on the QoE and other considered metrics.
3- The proposed communication steps on page 5 are not clear. Please elaborate on them by explaining the details.
4- The proposed approach used a Cubic algorithm. What about other congestion algorithms? Please clarify.
5- Please clarify how Eq. (4) is designed.
6- The performance evaluation part must be rewritten with different scenarios.
7- The manuscript should be revised for English grammar errors.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
This paper proposes an active bandwidth allocation framework based on congestion control, which can allocate bandwidth in a distributed manner and ensure the bandwidth utilization of links. The following issues are suggested to be revised.
- It is advisable to add a description of the conditions under which the results of Figure 1 are produced.
- If possible, please add relevant literature. In addition, other methods compared with the proposed method should explain their important differences in the relevant literature.
- According to the Eq. (2), it can be simplified to 1/n. Such a result is not Jain's fairness coefficient. Please check it carefully.
- Abbreviations mentioned for the first time in the article (eg. IPPO, QMix) should be presented with their full names.
- Symbols used in equations should be defined or explained. Such as J in Eq. 7, E and Clip in Eq. 8, and V in Eq. 10, these symbols need to be defined or explained. The ε is defined as the learning rate (Line 245). However, does ε in Eq. 8 (1-ε, 1+ε) also represent the learning rate?
- In Line 256-257 (in page 7), “ Based on … is shown in Table 1”. But, Table 1 (in page 10) is the experimental results of under dynamic bandwidth.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
The introduction provides a sufficient background and includes all relevant references.
Not all references cited are relevant to research. The paper, as for a publication planned in an academic journal, has a not very extensive list of references to literature. It is probably useful to have a deeper discussion on, for example, video artifact evaluation and video streaming service quality, also with different packet loss patterns.
The research design is appropriate.
Not all methods are adequately described. It is not clear what the authors sometimes mean by Quality of Experience (QoE). For example, in "Figure 1. The perceived quality of different videos at each bitrate.", The units of the vertical axis ("Experienced Quality") are not described at all. From reading the following paragraphs, it can be concluded that QoE means Video Multimethod Assessment Fusion (VMAF), however: firstly, it should be explicitly said from the beginning, and secondly, VMAF is only an objective, fully referential metric of video quality, meanwhile, the concept of QoE is much broader than the concept of video quality.
All results are clearly presented.
The conclusions are supported by the results.
Extensive editing of the English language and style is required. The manuscript should be carefully re-read to eliminate minor typos, for example missing spaces, like in "possible[2]" or "Bitrate(kbit/s)". Also, it is a good idea to have high-level headings separated from sub-level headings with short introductory paragraphs. Meanwhile, in the manuscript there are situations like: “5. Evaluation 289 5.1. Method 290 5.1.1. Evaluation Indicators 291”.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The authors carefully addressed my comments. I think the current manuscript can be considered for regular publication by the Future Internet Journal.