An Efficient Location-Based Forwarding Strategy for Named Data Networking and LEO Satellite Communications
, José Carlos López-Ardao 2,*, Miguel Rodríguez-Pérez 2, Sergio Herrería-Alonso 2, Andrés Suárez-González 2 and Raúl F. Rodríguez-Rubio 2
Round 1
Reviewer 1 Report
The submission on “A Routing-Less Forwarding Strategy for Named Data Networking and LEO Satellite Communications“ is proposing a location-based forwarding strategy for LEO satellite networks based on knowledge of the relative position of the satellites and the grid structure formed by the ISLs to perform the forwarding of NDN packets.
The work is addressing a new field of delivery of data via LEO networks and characterizes the arising general challenges. However, the own contribution of a forwarding strategy approach is restricted to 2 pages of text plus 2 large figures showing some special evaluation results on pages 4-6. With such a little contribution (where it is questionalbe how it can be further divided among 6 authors ?) it is not clear, whether the approach is an “excellent candidate …”. In general, the authors mention that low latency, bandwidth efficiency, and frequent handovers between moving LEOs and the comsumers/producers are main challenges to be solved, but most of those aspects are not mentioned or solved in the main part, except for some efficiency in bandwidth usage shown in Fig. 1. Therefore the authors should invest at least 2-3-fold effort for a submission and they should devote at least half of the space of a submission to describe own work instead of general describtions of a research area.
Remarks on details:
The term “Routing-Less” in the title is not to the point and should be removed, because NDN processing has to include routing tasks as well, even if they are organized different from “IP routing”. On the contrary, the authors often refer to “NDN routers”, “NDN routing protocols”, “router caches” etc. in their submission.
On page 1, the statement “This means that the NDN architecture allows content providers to keep copies of their most popular content in caches located close to the users.” sounds as if NDN has invented caching, but web caching is in use on the Internet since 30 years. The authors should compare NDN to state of the art technology, which is web caching via CDNs and clouds, and not to a situation 30 years ago without caches.
On page 5, please provide units (km, seconds ?) in the statements
“…, 400 altitude. The tracking period was established to 20.”
In the references:
Please remove repetition several times: "Proceedings of the Proceedings of the"
In reference 9. and 18., please add the publisher
In reference 17., please remove "ndnSim authors.", if authors are unknown, or insert the authors, if they are known.
Author Response
We would like to sincerely thank you for your helpful comments and suggestions about the former version of our article. Let us report the changes made in this new submission, in accordance with the recommendations of all the reviewers. Also note that we have rewritten some parts of the article to make the text more understandable and easier to follow. We do really hope that the changes made to the paper heed the suggestions of the reviewers and improve its quality so as to make it merit publication in your journal.
Response to comments
Point 1: The work is addressing a new field of delivery of data via LEO networks and characterizes the arising general challenges. However, the own contribution of a forwarding strategy approach is restricted to 2 pages of text plus 2 large figures showing some special evaluation results on pages 4-6. With such a little contribution (where it is questionable how it can be further divided among 6 authors ?) it is not clear whether the approach is an “excellent candidate …”.
Response 1: We are aware that six authors may appear excessive to undertake just the most-technical work in the paper. However, this paper is part of a much larger work on which the full team of authors is currently working. In any case, to address the concerns of the reviewer we have made explicit the individual contributions of every author in the “Author Contribution” section at the end of the paper. Regretfully, we had not properly indicated that in our previous version.
Point 2: In general, the authors mention that low latency, bandwidth efficiency, and frequent handovers between moving LEOs and the consumers/producers are main challenges to be solved, but most of those aspects are not mentioned or solved in the main part, except for some efficiency in bandwidth usage shown in Fig. 1. Therefore the authors should invest at least 2-3-fold effort for a submission and they should devote at least half of the space of a submission to describe own work instead of general descriptions of a research area.
Response 2: The paper has been significantly enhanced. On the one hand, a new section “Related Work” has been created, discussing existing works that address routing problems in LEO networks. On the other hand, we have added three new figures to clarify both the scenario and the forwarding strategy. Figure 1 depicts a portion of an example constellation showing the satellite orbits, the inter-satellite links and a possible path between two network locations. Additionally, Figure 2 shows the representation of the constellation network and its links as a grid of routing nodes. And Figure 3 has also been added to help to better understand how our proposed forwarding mechanism works. Finally, additional results and graphics have also been included and the conclusions section has been also enhanced.
Point 3: The term “Routing-Less” in the title is not to the point and should be removed, because NDN processing has to include routing tasks as well, even if they are organized differently from “IP routing”. On the contrary, the authors often refer to “NDN routers”, “NDN routing protocols”, “router caches” etc. in their submission.
Response 3: As suggested by the reviewer, we have removed the term “Routing-Less” from the title. We recognize that this term could be misunderstood, so we have replaced it with the more fitted term “Location-Based”.
Point 4: On page 1, the statement “This means that the NDN architecture allows content providers to keep copies of their most popular content in caches located close to the users.” sounds as if NDN has invented caching, but web caching is in use on the Internet since 30 years. The authors should compare NDN to state of the art technology, which is web caching via CDNs and clouds, and not to a situation 30 years ago without caches.
Response 4: As the reviewer correctly points out, this statement seems to indicate that NDN has introduced in-network caching on the Internet for the first time. Actually, we were just interested in highlighting that NDN enables in-network caching in an inherent way, so we have replaced it with the following sentence: “This means that the NDN architecture natively provides in-network caching without involving content providers.”
Point 5: On page 5, please provide units (km, seconds ?) in the statements “…, 400 altitude. The tracking period was established to 20.”
Response 5: We have added the corresponding units to those parameters.
Point 6: In the references: Please remove repetition several times: "Proceedings of the Proceedings of the". In reference 9. and 18., please add the publisher. In reference 17., please remove "ndnSim authors.", if authors are unknown, or insert the authors, if they are known.
Response 6: We have corrected all the highlighted errors in the bibliography.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The paper is about routing in ISL LEO networks. Although the research area is quite old, starting with Iridium and some other possible real constellations (e.g. Celestri, Teledesic, ...), the authors approach the problem from a different perspective and try to investigate and introduce Information Centric Networks in LEO ISL satellite communications.
The authors propose a new location-based forwarding strategy that can enable efficient and effective use of the NDN architecture in LEO satellite networks. The proposed forwarding strategy uses the knowledge of the relative position of the satellite nodes, together with the grid structure formed by the inter-satellite links to perform the forwarding of the NDN Interest packets while completely avoiding the use of routing protocols and the associated scaling problems.
The paper is well written, but I would suggest the following improvements:
1) The introduction could be expanded or you could add a section on related work where you highlight some related work, including older work, that addresses the routing/forwarding problems in ISL networks and then compare your approach to these "baselines" or "traditional" approaches. Please also address the following issue:
Since we know that the LEO ISL Inclined constellations are highly dynamic and that the links are not equally distant, conventional routing tends to route closer to the poles, thus causing congestions and consequently Oscillations, discussed in ("Oscillation suppression for traffic class dependent routing in ISL network, Svigelj, A., Mohorcic, M., Kandus, G. IEEE Transactions on Aerospace and Electronic Systems, 2007, 43(1), pp. 187-196). How does the constellation "properties" affect your procedure?
2) A graphical example (on the ISL grid) of the proposed forwarding strategy would be appreciated.
3) The authors should also spend some time discussing the reality of the simulation setup and the possible use of the system under real conditions. In particular, they should discuss how the Producers and Consumers are distributed, especially when considering Earth where the population is not evenly distributed.
4) You can add some Figure with the constellation that also shows the current positions of Producers/Consumers.
5) You can add some additional results e.g. histograms plots with Link usage.
Author Response
We would like to sincerely thank you for your helpful comments and suggestions about the former version of our article. Let us report the changes made in this new submission, in accordance with the recommendations of all the reviewers. Also note that we have rewritten some parts of the article to make the text more understandable and easier to follow. We do really hope that the changes made to the paper heed the suggestions of the reviewers and improve its quality so as to make it merit publication in your journal.
Response to comments
Point 1: The introduction could be expanded or you could add a section on related work where you highlight some related work, including older work, that addresses the routing/forwarding problems in ISL networks and then compare your approach to these "baselines" or "traditional" approaches.
Response 1: We have added a new section “Related Work” that discusses works addressing routing problems in LEO networks.
Point 2: Please also address the following issue: Since we know that the LEO ISL Inclined constellations are highly dynamic and that the links are not equally distant, conventional routing tends to route closer to the poles, thus causing congestions and consequently Oscillations, discussed in ("Oscillation suppression for traffic class dependent routing in ISL network, Svigelj, A., Mohorcic, M., Kandus, G. IEEE Transactions on Aerospace and Electronic Systems, 2007, 43(1), pp. 187-196). How does the constellation "properties" affect your procedure?
Response 2: This issue has been addressed in the new “Related work” section.
Point 3: A graphical example (on the ISL grid) of the proposed forwarding strategy would be appreciated.
Response 3: We have added three new figures to clarify both the scenario and the forwarding strategy. Figure 1 depicts a portion of an example constellation showing the satellite orbits, the inter-satellite links and a possible path between two network locations. Additionally, Figure 2 shows the representation of the constellation network and its links as a grid of routing nodes. Finally, Figure 3 has also been added to help to better understand how the mechanism works.
Point 4: The authors should also spend some time discussing the reality of the simulation setup and the possible use of the system under real conditions. In particular, they should discuss how the Producers and Consumers are distributed, especially when considering Earth where the population is not evenly distributed.
Response 4: We have added some text to the Results section explaining why the simulated ground nodes are arranged between latitudes 35º to 45º. In essence, we have tried to put the nodes in locations corresponding to densely populated areas of the globe.
Point 5: You can add some Figure with the constellation that also shows the current positions of Producers/Consumers.
Response 5: We believe that with the newly added Figure 1 and the more detailed description of the consumer and producers locations this is no longer needed.
Point 6: You can add some additional results e.g. histograms plots with Link usage.
Response 6: We have added additional results to illustrate the different amounts of traffic that can be carried by the network in different scenarios, even when considering the effect of the caches in the forwarding nodes, which had been neglected in our previous version.
Author Response File: Author Response.pdf
Reviewer 3 Report
The authors propose a named network for LEO satellite communication. The concept is very interesting. Comments are below.
(1) Authors need to conduct a more comprehensive literature review of LEO
(2) Comparision of the proposed method with IP V4 or V6 is needed.
(3) Results are very sparse. Please enrich your results.
(4) Communication vulnerability needs to be discussed.
(5) Data properties need to be extensively studied in the article as well.
Author Response
We would like to sincerely thank you for your helpful comments and suggestions about the former version of our article. Let us report the changes made in this new submission, in accordance with the recommendations of all the reviewers. Also note that we have rewritten some parts of the article to make the text more understandable and easier to follow. We do really hope that the changes made to the paper heed the suggestions of the reviewers and improve its quality so as to make it merit publication in your journal.
Response to comments
Point 1: Authors need to conduct a more comprehensive literature review of LEO
Response 1: We have added a new “Related Work” section that discusses works addressing routing problems in LEO networks.
Point 2: Comparison of the proposed method with IP V4 or V6 is needed.
Response 2: Our proposed forwarding strategy has been designed for NDN networks, an architecture paradigm quite different from the traditional TCP/IP architecture. NDN is a data-centric architecture while TCP/IP is host-centric. Additionally, our proposal takes advantage of the separation of forwarding concerns between the routing network layer and the strategy layer in NDN. In fact, we make use only of the latter. A direct correspondence with the architecture of IP does not exist.
In any case, we think that in the case of LEO satellite networks with TCP/IP architecture, an analogous strategy would perform very similarly as the case when we measured the performance of different requested prefixes for each consumer in NDN.
Point 3: Results are very sparse. Please enrich your results.
Response 3: We have added additional results and graphics.
Point 4: Communication vulnerability needs to be discussed.
Response 4: We have added a footnote to the Results section clarifying the cause of the packet losses measured during our experiments. It is due to handovers at the consumer location when there is/are pending Interest packets. In any case, the effects of these losses is minor and is orthogonal to the routing of Interest packets, which is the problem solved with the GeoTag-based forwarding strategy. We identify two possible solutions to fix/ameliorate this problem in the paper: a) use a more sophisticated mechanism to forward the Data packets back when near the consumer and b) rely on retransmissions if/while the loss percentage is low enough. In this paper we have opted for b) for simplicity.
Point 5: Data properties need to be extensively studied in the article as well.
Response 5: This forwarding strategy is suitable for any type of traffic as it distributes the traffic between two disjoint paths with a minimum number of hops, but it is clear that traffic with a multicast nature can benefit significantly of the underlying NDN network
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The submission on “A Routing-Less Forwarding Strategy for Named Data Networking and LEO Satellite Communications“ is proposing a location-based forwarding strategy for LEO satellite networks based on knowledge of the relative position of the satellites and the grid structure formed by the ISLs to perform the forwarding of NDN packets.
The work is addressing a new field of delivery of data via LEO networks and characterizes the arising challenges. Compared to a previous submission, the authors have substantially extended the Sections on their own work and results, such that a minimum amount of own work is now presented worthwhile for a publication. Results on load and throughput aspects are presented on a broader basis. The extended presentation has made the work more meaningful and attractive for readers.
Comments on details:
The usage of the wording of “routing” versus “forwarding” throughout the paper still seems to be inconsistent and/or unclear. The main part refers to “NDN routing in LEO satellite networks” or “adaptive routing protocols in large LEO satellite networks” or “routing Interest packets along a grid”, etc.
However, when finally the conclusion highlights
“the forwarding of the NDN Interest packets without the need of using routing protocols”
there is no concrete explanation what this should mean, i.e. how can you avoid to use any information about the network layer topology (either direct info or via an indirect background network information source)? Forwarding on the link layer is possible without routing on the network layer, e.g., in a fixed network topology on preconfigured static routes, but LEO satellite networks are not static? Therefore, please explain in more detail what the wording “forwarding without routing” should mean or rephrase.
The statement on page 1:
“This means that the NDN architecture natively provides in-network caching without involving content providers.”
has been updated, but it is devious to claim that caching works without cooperation of content providers. If Google does not like in-network caching, this will be blocked e.g. via encrypted application layer channels. Please rephrase the statement on a realistic background. A discussion of obstactes of caching in IP networks in a more realistic view can be found e.g. in Section 3, 3.3 in
G. Hasslinger and F. Hartleb, Content delivery and caching from a network provider’s perspective, Computer Networks 55 (2011) 3991-4006
Fig. 1 should be placed in Section 4 at the end of the page or on the next page.
On the y-axis of Fig. 6 and Fig. 7 the notation “(bytes)” should be removed, because the measure is obviously in %, not in bytes.
Author Response
Point 1: The usage of the wording of “routing” versus “forwarding” throughout the paper still seems to be inconsistent and/or unclear. The main part refers to “NDN routing in LEO satellite networks” or “adaptive routing protocols in large LEO satellite networks” or “routing Interest packets along a grid”, etc. However, when finally the conclusion highlights “the forwarding of the NDN Interest packets without the need of using routing protocols”, there is no concrete explanation what this should mean, i.e. how can you avoid to use any information about the network layer topology (either direct info or via an indirect background network information source)? Forwarding on the link layer is possible without routing on the network layer, e.g., in a fixed network topology on preconfigured static routes, but LEO satellite networks are not static? Therefore, please explain in more detail what the wording “forwarding without routing” should mean or rephrase.
Response 1: The sentence in the abstract and the introduction section (l. 76) “... without the need for any conventional routing protocol.” was replaced by a more clear sentence “So, forwarding at each node is done using only local information (node and destination locations), without the need of interchanging information between nodes, as is the case with conventional routing protocols.
The sentence (l. 108, l.172 and l.317) “... to completely avoid the use of adaptive routing protocols” was replaced by “... to completely avoid the need of interchanging information between nodes, as is the case with conventional adaptive routing protocols”
Point 2: The statement on page 1: “This means that the NDN architecture natively provides in-network caching without involving content providers.” has been updated, but it is devious to claim that caching works without cooperation of content providers. If Google does not like in-network caching, this will be blocked e.g. via encrypted application layer channels. Please rephrase the statement on a realistic background. A discussion of obstacles of caching in IP networks in a more realistic view can be found e.g. in Section 3, 3.3 in G. Hasslinger and F. Hartleb, Content delivery and caching from a network provider’s perspective, Computer Networks 55 (2011) 3991-4006
Response 2: It is true that any content producer can direct its clients to ask for different names even for the same content and thus avoid any in-network caching. The statement is not meant to imply that NDN forces network caching on producers, but to highlight that producers can benefit from network caching without having to fundamentally modify their applications. In the same vein, NDN network operators can choose not to cache some content if that seems appropriate from their business or operational point of view. In any case, we have rephrased the sentence so that it does not appear as something forced upon producers in this new version.
Point 3: Fig. 1 should be placed in Section 4 at the end of the page or on the next page. On the y-axis of Fig. 6 and Fig. 7 the notation “(bytes)” should be removed, because the measure is obviously in %, not in bytes.
Response 3: Done.
A PDF with the differences is attached
Author Response File: Author Response.pdf
Reviewer 3 Report
Authors have addressed my comments and therefore I recommend for publication.
Author Response
We would like to sincerely thank you for your helpful comments and suggestions
