======= Review 1 =======

*** Familiarity: Rate your familiarity with the topic of the paper.
Familiar (I am well aware of research work in this topic) (3)

*** Recommendation: Your overall rating
Accept if room (top 30% but not top 20%, borderline) (3)

*** Contributions: What are the major issues addressed in the paper? Do you consider them important? Comment on the novelty, creativity, impact, and technical depth in the paper.

This paper introduces the routing problem under multiple QoS constraints for overlay network, which is an important topic, and gives a new routing protocol based on existing Table-driven and Demand-driven routing protocols. This paper takes QoS constraints into routing considerations. It is shown in the simulation that the proposed method in this paper makes improvements in some aspects including convergence of the algorithm and message overhead from existing methods.

*** Strengths: What are the major reasons to accept the paper?

This paper has good organizations and shows a clear architecture for the process of the step-by-step design of the routing protocol. Examples given in this paper help state the proposed method clearly. Extensive simulations are carried out to show the effectiveness of the proposed method.

*** Weaknesses: What are the major reasons NOT to accept the paper?

The proposed method in this paper is based on both table-driven and demand-driven protocols. However, table-driven method is weak in the changing topology and demand-driven method is weak in the case of increasing number of source to destination pairs. This paper does not give sufficient discussions on the drawbacks of these two aspects. It is confused that whether the proposed method is robust to the changing topology and increasing number of source to destination pairs. Under practical applications, an overlay network may suffer topology changes and the number of sources may vary in a wide range. Hence the paper should take these factors into design considerations.

*** Detailed Comments: Please provide detailed comments that will help the TPC assess the paper and help provide feedback to the authors.

This paper solves an important problem in QoS constrained overlay networks with a routing method based on existing table-driven and on demand routing protocols. It is shown in the simulation results that the method converges in a few steps for a static environment. Comparisons are carried out to show the improvements of the proposed method. However, there are several problems needed to be addressed.

1. The proposed method is for real-time application. It only considers the propagation delay, i.e. the delay of the link in this paper. Most existing Table-driven routing protocols (such DSDV) and on-demand protocols (such as AODV) do not provide real-time guarantees. The proposed method borrows ideas of those protocols but does not make any improvement to ensure end-to-end delay in real-time applications.

2. It is strange that there is a system implementation section in the end of this paper but no result is shown on the testbed with the proposed method. Readers may get confused with the real function of the implementation.

3. As mentioned above, the proposed method in this paper is based on both table-driven and demand-driven protocols. However, table-driven method is weak in the changing topology and demand-driven method is weak in the case of increasing number of source to destination pairs. This paper does not give sufficient discussions on the drawbacks of these two aspects. It is confused that whether the proposed method is robust to the changing topology and increasing number of source to destination pairs.

======= Review 2 =======

*** Familiarity: Rate your familiarity with the topic of the paper.
Some knowledge (I am marginally aware of research work in this topic) (2)

*** Recommendation: Your overall rating
Likely accept (top 20% but not top 10%, significant contribution) (4)

*** Contributions: What are the major issues addressed in the paper? Do you consider them important? Comment on the novelty, creativity, impact, and technical depth in the paper.

This paper provides methods to meet individual QoS metrics in multiple QoS constraints environments. Instead of using a composite single metric for multiple QoS constraints, this paper disseminates individual QoS metric information, and end-to-end paths are set up based on the disseminated information. As a result, this paper reduces both false positive and negative rates compared to existing techniques. The problem is not new but well addressed. Technical depth is also satisfactory.

*** Strengths: What are the major reasons to accept the paper?

This paper is well structured and presented. The importance of considering multiple QoS constraints and limitations of existing work are clearly presented. The proposed methods are clear and technically satisfactory. Evaluation is well organized with diverse perspectives including overhead, stabilization times, false rates, etc, supporting validity of the proposed methods.

*** Weaknesses: What are the major reasons NOT to accept the paper?

Of the experimental results, violation rates in a dynamic environment are interesting. To make it more practical, it can be done more in dynamic environments: in their experiments, the heaviest environment is 5 connections/sec, which is fairly light in reality.

*** Detailed Comments: Please provide detailed comments that will help the TPC assess the paper and help provide feedback to the authors.

This paper presents methods to serve applications with multiple QoS constraints in a service overlay network. Unlike existing techniques for multiple QoS constraints using a single composite metric, this paper attempts to meet individual QoS constraints by considering individual QoS requirements in path discovery. For this, the proposed methods include a path vector protocol to disseminate the best path for each QoS metric, so that each node can obtain end-to-end reachability information. Sometimes the path vector protocol canÂ’t find a path even though there exist some paths satisfying the constraints because it gives only the best path information. In this case, on-demand route discovery is performed to search any paths by querying to neighbors. In addition, path adaptation is provided to handle network dynamism.

Overall, this paper is well presented and contains extensive results. Motivations and related work are adequately described, and the authorsÂ’ approach is clearly stated. The algorithm to find an end-to-end path is well described to show the overall mechanism of the finding. The protocol flow chart is valid to show overall procedures of the suggested methods. Evaluation is well organized and the results support the proposed methods.

======= Review 3 =======

*** Familiarity: Rate your familiarity with the topic of the paper.
Familiar (I am well aware of research work in this topic) (3)

*** Recommendation: Your overall rating
Accept if room (top 30% but not top 20%, borderline) (3)

*** Contributions: What are the major issues addressed in the paper? Do you consider them important? Comment on the novelty, creativity, impact, and technical depth in the paper.

This paper studies routing subject to multiple QoS constraints. It is a good comprehensive study, but lacks clear cut contributions.

*** Strengths: What are the major reasons to accept the paper?

Nice problem to study.

*** Weaknesses: What are the major reasons NOT to accept the paper?

The paper lacks clear cut contributions. While I was reading this paper, I kept think "what is the contribution"?

*** Detailed Comments: Please provide detailed comments that will help the TPC assess the paper and help provide feedback to the authors.

The paper has insightful ideas on multiconstrained routing. However, it lacks clear cut contributions.