TCP-LoRaD: A Loss Recovery and Differentiation Algorithm for Improving TCP Performance over MANETs in Noisy Channels

Round 1

Reviewer 1 Report

The authors in this work improve TCP by proposed a method TCP-LoRaD that use the main parameter of the TCP. 

Some remarks should be clarified:

1- The figures are not clear for the readers

2- In equation (7), how to learn the value of alpha 

3- How to choose or estimate the noise channel. 

Author Response

Please see attached document.

Author Response File: Author Response.pdf

Reviewer 2 Report

The authors present a novel TCP protocol for lossy networks. The paper is already substantial in size.

1. Nevertheless, pls introduce loss recovery mechanisms for selected protocols (e.g., TCP WELCOME)

2. Pls, provide an enhanced rationale for the selected parameters. This would allow the reader to, for example, understand why in the congested situation RTTc > RTT. It would be good to provide insights into the expected values of those parameters (eq. 1-11) in some hypothetical cases, allowing the reader to easily follow the paper's key message. 

3. Pls, expand ALL abbreviations, e.g., ACK, MSS, FTP. 

4. Use standard SI units, e.g., watt -> W, meters/second -> m/s

5. The "increasing/varying node" comparison should always mention the increasing/varying NUMBER of nodes. 

6. Pls, provide some statistical information about all presented figures, i.e., how many trials per point are provided. Pls, display other parameters such as min, max, and standard deviation in all figures. 

7. In Fig 5, pls explain why there is a substantial difference in the case of 5 nodes, while in other situations, i.e., 10-20 nodes, the difference is so minor.

8. In Fig. 6, the same for 20 nodes against 5-15 nodes.

9. In Fig. 13, pls comment on the MSS/MTU in this network.

10. Pls compare fairness (e.g., the Jain index) between TCP-LoRaD and other TCP versions (e.g., TCP Reno) for selected situations, e.g., upon a good channel.

11. The paper claims to implement TCP-LoRaD in Riverbed Modeler (i.e., OPNET). However, it mentions that TCP-WELCOME was implemented in NS-2. Therefore, what implementation of TCP-WELCOME is used in this paper? Is it TCP-WELCOME ported from NS-2? If so, can you point a link, e.g., Github implementation?

12. If possible, pls provide the implementation of TCP-LoRaD through Git or other channels. 

Author Response

Please see attached document.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

The paper is good to go.