Final acceptance letter: The paper has been quickly re-evaluated on behalf of the managing editor in this 2nd revision, without a new round or peer review, based on the fact that the required revisions were minor and easy to verify. Also based on the exhaustive accompanying letter provided by authors, which properly explains the revisions made, it appears that authors have now completely satisifed the final minor changes required by reviewers, and for this reason the paper can now be accepted. -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Reviews Round 2: [Note from the managing editor: The reviewers agree ont he achieved quality of the paper after the first round of revision. On the other hand, one reviewer (#4) still requires a minor change and clarification which could be easily made in the paper. For this reason, the paper decision is to allow a (very) minor revision of the paper which will be evaluated under the responsibility of the managing editor only (that is, not another round of review). By the way, authors are recommended to evaluate the inclusion of citations of the following recent papers, which appears to be relevant references to complete the state of the art analysis for this paper contribution: M. Di Felice, L. Bedogni, L. Bononi, "Group Communication on Highways: An Evaluation Study of Geocast Protocols and Applications", Elsevier's Ad Hoc Networks Journal, vol.11, No.3, pp. 818-832, May 2013, A.J. Ghandour, M. Di Felice, H. Artail, L. Bononi, "Dissemination of Safety Messages in IEEE 802.11p/WAVE Vehicular Network: Analytical Study and Protocol Enhancements", Elsevier's Pervasive and Mobile Computing (PMC) Journal, http://dx.doi.org/10.1016/j.pmcj.2013.03.003 ] Reviewer #2: I am satisfied with this revisioned paper. It can be accepted Reviewer #3: This paper studies the performance of the back-off mechanism and the role of the congestion window on the control channel of a vehicular networks, specially focusing on high density scenarios. The authors show that the broadcast and periodic nature of safety messages have an important impact on the optimal value of the contention windows. Then, basing on these findings, they propose a reverse back-off mechanism to rule the transmission of CAM and DEMN messages in vehicular networks. Personally I found this paper really interesting: it is well written and it clearly explains a new mechanism that should have margin to improve the communication in vehicular networks. The underlying idea is quite simple, but it is proved through a precise mathematical analysis and an extensive simulation campaign. Also, I appreciated the effort the authors put in adopting realistic propagation and mobility models for the simulations. Moreover, the authors provided exhaustive answers to the reviewers' comments and they fixed all the doubts emerged in the previous version of this manuscript. Indeed, I think that this paper is ready to be published in this journal. Reviewer #4: I would like to thank the authors for answering my concerns adequately. I still have the following question: -Earlier, I mentioned that dropping expired messages is not a requirement of WAVE: "Actually, The approach described by the author is one possible option for queue handling. However, this is a chipset implementation issue not specified in the WAVE standard. Alternative options are possible and might be adopted where the old beacon is not flushed" The authors mentioned in their reply that ETSI ITS "clearly required dropping expired messages" but they did not discuss WAVE requirements. First, please make it clear in Section 2.1 that this is an ETSI requirement. Second, what about WAVE compatibility when a dedicated radio is used to monitor CCH? Can we assume that the suggested work is only compatible with the European standard? Please elaborate on this. -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Reviews Round 1: [Note from the managing editor: the paper has received 3 reviews from expert reviewers indicating the merit of the paper and some directions for improvement leading to a decision of a MAJOR revision required (and re-review). Authors are invited to follow the reviewers' recommendations in the revision of the paper contribution, and to resubmit the revised paper within 90 days. Please note that a detailed letter must be provided with a detailed response to the reviewers' comments, including the defense of the paper and the details of changes made in response to reviewers' comments, to ease the re-review process.] Reviewer #2: This paper has a potential to be accepted, but some important points have to be clarified or fixed before we can proceed and a positive action can be taken. We here summarize this points: 1. It is really unclear to me the vehicular scenario that the authors take into consideration. This is an important point before we can start to reason about the proposed method seriously. We need to understand precisely if the proposed mechanism is able to face with the following: - different vehicles may have different transmission ranges and receiver sensitivities depending on the hardware devices they mount, - transmission ranges between a pair of vehicles can be asymmetric: that is, it may happen that vehicle a can receive packets from vehicle b, but not vice versa, - all this may be exacerbated by the fact that, with the passage of time, transmission ranges can vary due to a number of reasons, such as intervening obstacles and physical propagation effects (including multi-path and fading). 2. The authors seem to disregard or neglect some important results that have been recently achieved in this specific field. For example authors should not ignore the following: A. Amoroso, G. Marfia, M. Roccetti, ``Going Realistic and Optimal: A Distributed Multi-Hop Broadcast Algorithm for Vehicular Safety'', Computer Networks, Elsevier, vol. 55, n. 10, July 2011, pp. 2504-2519. We need to understand how the proposed approach is related with this result. 3. Authors should revise better and more the current literature in the field. For example a good survey can be found here: M. Di Felice, L. Bedogni, L. Bononi, "Group Communication on Highways: An Evaluation Study of Geocast Protocols and Applications", 2013 Elsevier Ad Hoc Networks. 4. I have many concerns about the style of this paper. I think that if the authors wish this paper is well considered by experts in the vehicular field, more attention should be devoted to discuss the application scenario. In essence the paper is too long and boring with a large use of mathematics that is often used without without clarity and sufficient motivations. I suggest to simplify it or better explain with realistic examples. 5. The provided simulative results are not completely convincing to me. Again they are too vague and generic. Again I suggest to take inspiration from the following paper where real on-field experiments were carried out. Provided the above questions are answered and problems are fixed, the paper can be reconsidered for publication. Reviewer #3: In this paper, the authors investigate the compatibility of the IEEE 802.11 medium access control protocol with the requirements of safety vehicular applications. Using an analytical framework, they study the performance of the back-off mechanism and the role of the contention window on the control channel of a vehicular network. Based on these findings, they propose a reverse back-off mechanism, specifically designed for road safety applications. The efficiency of the proposed enhancement scheme is proved through a simulation campaign. The overall level of the paper is good: even if it is quite simple, it is well written and some important considerations are highlighted. Moreover, the proposed mechanism seams to be an enhancement for high-density vehicular networks. In the following, there is a list of questions that the authors should answer: 1) The system model is trivial because it is suitable only for linear vehicular networks. Is it possible to extend the model for, at least, 2-D networks? 2) The hypothesis that at most 2 nodes can be involved in a collision is strong, in particular in case of high-density vehicular networks. It should be better try to relax it. 3) The authors should explain better the curves in Figures 3, 4 and 5. It is not clear why the curves are so different passing from 125 to 200 vehicles/km. 4) The caption of Figure 5 is wrong. It does not concern 160 vehicles/km only. 5) In section 5.2 the authors describe the simulation scenario. In particular, they specify that the statistics are computed using the final 5 minutes of the simulation, when the system already reached a steady-state. It is not clear if the authors are talking about the simulated or the real time. If they intended the simulated time, the simulation risks to be too short to capture the dynamic of a vehicular network. Reviewer #4: -The paper entitled "Reverse Back-off Mechanism for Safety Vehicular Ad Hoc Networks" studies the performance of the 802.11p MAC protocol for vehicular environment through an analytical model. It also proposes a new back-off mechanism that is shown using simulation study to result in better performance in vehicular environment. -The paper is well written. The Introduction and Background sections provide useful information for the readers. Nevertheless some information presented is not accurate. For example, in Section 2.1, the authors assume when a new beacon is available while a previous one is still waiting in the queue, the old beacon needs to be dropped. This assumption is used later at the end of Section 2 as an argument for not adopting oCSMA. Actually, The approach described by the author is one possible option for queue handling. However, this is a chipset implementation issue not specified in the WAVE standard. Alternative options are possible and might be adopted where the old beacon is not flushed. -Moreover, in Section 2.2, the authors argue that low speed vehicle is not necessarily due to increased road traffic. However, adaptive beaconing frequency schemes based on velocity assume that vehicles running at low speed do not need a high frequency of status update. Also, the authors relate beaconing frequency reduction only to the vehicle speed. However, several schemes were proposed where beaconing frequency is made a factor of different parameters such as channel contention. -In Section 3, the authors mention only two references for analytical model in VANETs. Several studies were omitted such as: C Campolo, A Molinaro, A Vinel and Y Zhang, "Modeling Prioritized Broadcasting in Multichannel Vehicular Networks", IEEE Transactions on Vehicular Technology 61 (2), 687-701. A. J. Ghandour, M. Di Felice, H. Artail and L. Bononi, "Dissemination of Safety Messages in IEEE 802.11p/WAVE Vehicular Network: Analytical Study and Protocol Enhancements", Pervasive and Mobile Computing Elsevier, http://dx.doi.org/10.1016/j.pmcj.2013.03.003 J. Misic, G. Badawy and V. Misic, "Performance Characterization for IEEE 802.11p Network with Single Channel Devices", IEEE Transactions on Vehicular Technology, 68(4), pp. 1775-1789, 2011. -A major assumption made in the manuscript is that number of nodes involved in a collision is two. This assumption needs to be verified especially in high density nodes where collision between multiple nodes is highly probable at low values of contention window. -The slot definition in the proposed model is different from 802.11p which results in large confusion for the reader. The slot value is set to 66.7us in Section 3.5 without justification. This confusion yields to an ambiguity in the definition of "duration of a beacon in slots N_s". -A major concern about the proposed reverse back-off mechanism is its performance in low to medium traffic scenarios. These scenarios are the dominant on the road. In the proposed scheme, whenever the vehicle senses the channel to be busy and enters back-off, a high delay will be incurred prior to the beacon transmission. This delay might be in the order of several milliseconds in a low to medium traffic scenario which results in a low channel utilization. -Several studies such as the one below present enhancement to the existent back-off mechanism. It is important to discuss these studies and compare their performance with the "reverse back-off scheme". M. Di Felice, A. J. Ghandour, H. Artail and L. Bononi, "On the Impact of Multi-channel Technology on Safety-Message Delivery in IEEE 802.11p/1609.4 Vehicular Networks", 21st IEEE International Conference on Computer Communication Networks (ICCCN 2012), July 30 - August 2, 2012, Munich, Germany. -Receiver Sensitivity and SINR values in Table 1 are not compatible with the literature. *********************************************************************************************************************************************************************************************** Comments for previous version submitted at IEEE Transactions on Mobile Computing *********************************************************************************************************************************************************************************************** Editor Comments to the Author: Based on the recommendations of the 3 attached reviews (1 reject and 2 Revise and resubmit as new) and my own evaluation of the paper, unfortunately, the manuscript cannot be accepted for publication to TMC in its current form. The overall recommendation for this paper is "Reject". According to the 1st reviewer, although the manuscript is very well-written, it studies a well-known problem to which many different solutions/approaches/proposals exist. In addition, contributions are marginal. Reviewer #2 states that the paper provides some significant advancement to the state-of-the-art, yet, has concerns regarding the correctness of the present theoretical analysis of beaconing performance. Furthermore, the analysis seems superficial from a technical viewpoint. According to Reviewer #3, the authors propose a relatively new method of contention window management in VANETs, however, its main contribution is a simple heuristic. The analysis in the paper is very basic probabilistic analysis mainly based on prior work. Also, the main contribution, i.e., reverse BEB, is based on observations only and is neither based on a new analytical approach nor related to aforementioned analysis. *********************** Reviewer Comments Please note that some reviewers may have included additional comments in a separate file. If a review contains the note “see the attached file” under Section III A – Public Comments, you will need to log on to ScholarOne Manuscripts to view the file. After logging in, select the Author Center, click on the “Manuscripts with Decisions” queue and then clicking on the “view decision letter” link for this manuscript. You must scroll down to the very bottom of the letter to see the file(s), if any. This will open the file that the reviewer(s) or the associate editor included for you along with their review. Reviewer: 1 Recommendation: Revise and resubmit as “new” Comments: The manuscript is very well written, the authors have understood the general scalability problem in VANETs and are aware of related work in this context. However, I have several problems with the technical contributions of this paper, in particular: - The default fixed CW size is 15 for safety-related beacons (it doesnt matter what 802.11p says, the profile standards of ETSI are relevant here), hence, the benefit of an increased CW size reduces to at most 2 percentage points (from approx. 8 percentage points when comparing against a fixed CW size of 7) - cf. Figure 9 of the manuscript. - Assuming the maximum transmit power and a beaconing rate of 10 Hz in very high densed scenarios looks artificial to me (especially if the inter-vehicle distance is only 23m). However, in these extreme scenarios the illustrated benefit reaches its maximum. - In Figure 10, the authors show that an adaptive CW size is less than 1 percentage point better that an optimal fixed CW size. My conclusion from this figure is therefore: just increase the initial CW size for beacons and do not perform any adaptation. - In Figure 11 the authors compare the performance increase with respect to the number of consecutive beacon losses and claim that their reverse backoff mechanism reduces the probability that more than 20 beacons are lost in a row by 40% (not percentage points!). At the same time, they state that the probability that up to 10 beacons are lost in a row is increased by 20% (again, no percentage points!), which is even an argument against reverse backoff. While these values sound bigger than 1-2 percentage points: they are misleading. In particular: how often do these cases occur overall? From my own studies I know that this is not very often the case. To summarize the above: there is no significant technical contribution, nor an advancement that would justify the acceptance of this manuscript. It's a solid work, and I am sure that a lof of effort went into this study, but the way it is presented does not justify acceptance (it would be better not to propose a new algorithm/mechanism, but to raise the question whether a contention window adaptation provides significant benefits. then the authors can conclude with the apparent result that the benefit is only marginal). Additional Questions: 1. Which category describes this manuscript?: Research/Technology 2. How relevant is this manuscript to the readers of this periodical? Please explain under Public Comments below. : Relevant 1. Please explain how this manuscript advances this field of research and/or contributes something new to the literature. : The paper studies the role and impact of the contention window (CW) size in 802.11p on the performance of safety-related beaconing. The authors argue that too small CW sizes (i.e. 802.11p uses only 15 backoff slots per default) lead to an increasing number of packet collisions, hence, the backoff algorithm does not fulfill its objective. According to their analysis and simulation results, a higher CW size would be better in densed traffic situations. Since the traditional exponential backoff is not applicable to broadcast communications (it is not possible to detect a collision), the authors propose to start with relatively large CW sizes, and to decrease it if beacon transmissions time out locally, i.e. if they are not sent out. From my point of view, the manuscript studies a well-known problem to which many different solutions/approaches/proposals exist: scalability in vehicular communications. The corresponding related work is fortunately quite complete (they refer to recent surveys and articles) if only the general congestion control approaches are considered. Their solution/proposal however is not convincing to me: the benefit is in most of the cases only marginal (up to 2% higher packet reception rates in comparison to vanilla 802.11p), and the cases in which higher benefits are claimed are artificial (i.e. not realistic to me), for instance 43 vehicles/km/lane and a 10 Hz beaconing rate are contradicting. So how does this manuscript advance the field of research? In my opinion not very much. The contributions are marginal, adding only a slight benefit, although I have to say that the manuscript itself is very well written - the authors know what they are writing about. 2. Is the manuscript technically sound? Please explain under Public Comments below. : Yes 1. Are the title, abstract, and keywords appropriate? Please explain under Public Comments below. : Yes 2. Does the manuscript contain sufficient and appropriate references? Please explain under Public Comments below. : References are sufficient and appropriate 3. Does the introduction state the objectives of the manuscript in terms that encourage the reader to read on? Please explain under Public Comments below. : Yes 4. How would you rate the organization of the manuscript? Is it focused? Is the length appropriate for the topic? Please explain under Public Comments below. : Satisfactory 5. Please rate the readability of this manuscript. Please explain your rating under Public Comments below. : Easy to read 6. Should the supplemental material be included? (Click on the Supplementary Files icon to view files): Does not apply, no supplementary files included 7. If yes to 6, should it be accepted: After revisions. Please include explanation under Public Comments below. 8. If this manuscript is an extended version of a conference publication, does it offer substantive novel contributions beyond those of the previously published work(s)- i.e. expansion of key ideas, examples, elaborations etc. *New results are not required*: Yes Please rate the manuscript. Please explain under Public Comments below. : Good Reviewer: 2 Recommendation: Revise and resubmit as “new” Comments: The paper addresses an important topic, and provide some significant advancement to the state-of-the-art. However, several issues have to be solved before the paper can be considered for publication. A first major issue regards the correctness of the present theoretical analysis of beaconing performance. In particular, the authors mix different assumptions about vehicle distribution in the analysis, which is not acceptable. At the beginning of the analysis, although this is never explicitly mentioned in the text, the authors assume Poisson density of vehicles. Later on in the derivation, the authors assume instead vehicles distributed uniformly at random. While the two distributions are definitely related, they are not equivalent. Actually, they are profoundly different in probabilistic terms: with uniform distribution, the number of nodes (vehicles) is known in advance, and it can hence be directly controlled; furthermore, the event "x vehicles are located in area A" *is not* independent of event "y vehicles are located in area B", where A\intersect B = EmptySet, since the fact that x vehicles are in A influences the number of vehicles which can be located in B. Conversely, in Poisson distribution, the number of vehicles in an area cannot be directly controlled, but only its *expected number* can. Furthermore, the two events above are independent under Poisson distribution, since the total number of deployed nodes is not fixed as with uniform distribution. Given the above, it is clear that is not acceptable from a mathematical viewpoint to assume Poisson distribution early in the analysis, and uniform distribution later on. The authors should fix this important technical flaw for the paper to be considered for publication. More in general, the analysis seems superficial from a technical viewpoint. For instance, the fact that P_b can be expressed as in (2) derives from the unstated assumption that the busy slot probability can be modeled as an independent Bernoulli trial. Otherwise, the fact that we have N_T slots and that the expected number of busy slots is E[N_b] does not imply (2). Furthermore, in several derivations the authors use the standard technique of conditioning on independent, disjoint events, but this is not stated in the text. Another problem that I see is that the analysis is not at all related with the evaluation. There is no way of understanding, e.g., whether the presented analysis can be used to predict performance of an actual backoff strategy. This is a very important assessment for an analytical framework, which would otherwise be useless. The notion of network load should be formally defined: what does a network load of, say, 1.25 means? This is not explained in the text. Judging from what reported in Fig. 11, it seems that the proposed backoff strategy is not very reactive in reducing consecutive beacon unsuccessful tx: only when >10 beacons are lost the strategy performs better than the best non-adaptive backoff strategy. Is this actually the case? Please, clarify. Minor comments: - The caption of Fig 4 is wrong (it is a repetition of the caption of Fig 3). - The caption of Fig 5 mentions n_c=160, which is not in accordance with what reported in the plot. - the explanation of why the relationship between CW size and network load is opposite w.r.t. Bianchi's model should be moved from pag.11 to the end of Section 3, where it is first mentioned. Additional Questions: 1. Which category describes this manuscript?: Research/Technology 2. How relevant is this manuscript to the readers of this periodical? Please explain under Public Comments below. : Relevant 1. Please explain how this manuscript advances this field of research and/or contributes something new to the literature. : The paper presents an analytical model to study the performance of beaconing in vehicular networks. Although simplistic under some respects, the model aims at modeling distinguishing features of vehicular network, such as broadcast communications and periodic beacon transmission. Based on the insights derived from the analysis, the authors design a backoff mechanism aimed at dynamically adjusting CW size based on the number of expired beacons. The modified backoff mechanism is shown through simulation to provide a certain benefit with respect to standard 802.11p. 2. Is the manuscript technically sound? Please explain under Public Comments below. : Partially 1. Are the title, abstract, and keywords appropriate? Please explain under Public Comments below. : Yes 2. Does the manuscript contain sufficient and appropriate references? Please explain under Public Comments below. : References are sufficient and appropriate 3. Does the introduction state the objectives of the manuscript in terms that encourage the reader to read on? Please explain under Public Comments below. : Yes 4. How would you rate the organization of the manuscript? Is it focused? Is the length appropriate for the topic? Please explain under Public Comments below. : Satisfactory 5. Please rate the readability of this manuscript. Please explain your rating under Public Comments below. : Easy to read 6. Should the supplemental material be included? (Click on the Supplementary Files icon to view files): Does not apply, no supplementary files included 7. If yes to 6, should it be accepted: 8. If this manuscript is an extended version of a conference publication, does it offer substantive novel contributions beyond those of the previously published work(s)- i.e. expansion of key ideas, examples, elaborations etc. *New results are not required*: Yes Please rate the manuscript. Please explain under Public Comments below. : Fair Reviewer: 3 Recommendation: Reject Comments: In this paper, the authors propose a relatively new method of contention window management that aims to deliver messages in VANETs based on the number of collisions experienced. The algorithm makes the node more aggressive as collisions persist and urgent messages cannot be delivered. The authors provide some lengthy discussions on the effects of CW management techniques on MAC performance under 802.11 family of protocols. There are two main shortcomings of the work: 1- The analysis provided in the paper is very basic probabilistic analysis mainly based on prior work. By itself, it does not increase the value of the work. 2- The main contribution, i.e., reverse BEB, is based on observations only and is neither based on a new analytical approach nor related to aforementioned analysis. As a result, the paper's main contribution is a simple heuristic. Hundreds of such heuristics have been proposed in the literature for different purposes, and tens of them refer to VANETs. As such, the paper is not able to break away from the existing strain of works. Since analysis provided in the paper has no direct relation to the proposed heuristic, the analysis should not be considered as part of the work anyways. Therefore, the proposed work is nothing more than YAPA MAC protocol with no analytical guidance. The authors are encouraged to retrace their work and provide a relevant analysis of their proposal and incorporate such results into their algorithm design (and not only be inspired by the results of some tangential analysis). Additional Questions: 1. Which category describes this manuscript?: Research/Technology 2. How relevant is this manuscript to the readers of this periodical? Please explain under Public Comments below. : Relevant 1. Please explain how this manuscript advances this field of research and/or contributes something new to the literature. : The presented material introduces a relatively new method to manage contention window selection for VANETs, which has not necessarily been proposed in the past in the context of deadlines. 2. Is the manuscript technically sound? Please explain under Public Comments below. : Yes 1. Are the title, abstract, and keywords appropriate? Please explain under Public Comments below. : Yes 2. Does the manuscript contain sufficient and appropriate references? Please explain under Public Comments below. : Important references are missing; more references are needed 3. Does the introduction state the objectives of the manuscript in terms that encourage the reader to read on? Please explain under Public Comments below. : Could be improved 4. How would you rate the organization of the manuscript? Is it focused? Is the length appropriate for the topic? Please explain under Public Comments below. : Could be improved 5. Please rate the readability of this manuscript. Please explain your rating under Public Comments below. : Readable - but requires some effort to understand 6. Should the supplemental material be included? (Click on the Supplementary Files icon to view files): Does not apply, no supplementary files included 7. If yes to 6, should it be accepted: 8. If this manuscript is an extended version of a conference publication, does it offer substantive novel contributions beyond those of the previously published work(s)- i.e. expansion of key ideas, examples, elaborations etc. *New results are not required*: Yes Please rate the manuscript. Please explain under Public Comments below. : Poor