===== Review ===== *** Raw Score for (1) Relevance/Timeliness/Utility (Raw score for criterion (1) Relevance/Timeliness/Utility (appropriateness for the conference and symposium) Rate the manuscript on this criterion with a raw score from 1 to 5. A total calculated score will have a maximum value of five (5) and will be calculated from the four scored criteria using a weighting factor of 0.2, 0.3, 0.3, and 0.2, respectively. The weighing factor for this category is 0.2.): 5: Excellent 4: Superior 3: Satisfactory 2: Poor 1: Unacceptable Evaluation=Satisfactory (3) *** Raw Score for criterion (2) Technical Contribution (Raw Score for criterion (2) Technical Contribution (content, logic, development, etc.) Rate the manuscript on this criterion with a raw score from 1 to 5. A total calculated score will have a maximum value of five (5) and will be calculated from the four scored criteria using a weighting factor of 0.2, 0.3, 0.3, and 0.2, respectively. The weighing factor for this category is 0.3.): 5: Excellent 4: Superior 3: Satisfactory 2: Poor 1: Unacceptable Evaluation=Superior (4) *** Raw Score for criterion (3) Quality of Content (Raw Score for criterion (3) Quality of Content (originality, novelty, completeness, references, etc.) Rate the manuscript on this criterion with a raw score from 1 to 5. A total calculated score will have a maximum value of five (5) and will be calculated from the four scored criteria using a weighting factor of 0.2, 0.3, 0.3, and 0.2, respectively. The weighing factor for this category is 0.3.): 5: Excellent 4: Superior 3: Satisfactory 2: Poor 1: Unacceptable Evaluation=Superior (4) *** Raw Score for criterion (4) Quality of Presentation (Raw Score for criterion (4) Quality of Presentation (organization, grammar, language, graphics, etc.) Rate the manuscript on this criterion with a raw score from 1 to 5. A total calculated score will have a maximum value of five (5) and will be calculated from the four scored criteria using a weighting factor of 0.2, 0.3, 0.3, and 0.2, respectively. The weighing factor for this category is 0.2.): 5: Excellent 4: Superior 3: Satisfactory 2: Poor 1: Unacceptable Evaluation=Superior (4) *** (1) Relevance/Timeliness/Utility (Appropriateness for the conference and symposium Comments on the manuscript related to this criterion): The topic of the paper is very appropriate for the conference. *** (2) Technical Contribution (Content, logic, development, etc. Comments on the manuscript related to this criterion): The paper logic flows very smoothly from describing the reactor pattern to the analysis of the response time. The VPN example at the end showed the appicability of the approach to a communication application. *** (3) Quality of Content (Originality, novelty, completeness, references, etc. Comments on the manuscript related to this criterion): It is very dificult for me to comment on this topic area since I am not an expert in the domain of software performance. I am not familiar with the mechanisms that the paper used to estimate the response times. *** (4) Quality of Presentation (Organization, grammar, language, graphics, etc. Comments on the manuscript related to this criterion.): The presentation of the paper is very good. It requires though a good understanding of the SRN technique to appreciate its value. The authors might want to justify the intermediate states in the SRN model of the reactor pattern. For example I was not that celar why the SN1 & SN2 states are required. Also a brief description of what the different SRN symbols signify, i.e. a circle, a line, etc. Swapna: It seems like the reviewer did not quite get the SRN mechanism. I believe that we have explained what the transitions Sn1 and Sn2 accomplish. Also, due to the page restrictions, there is no room to elaborate on the graphical representation of SRN. ===== Review ===== *** Raw Score for (1) Relevance/Timeliness/Utility (Raw score for criterion (1) Relevance/Timeliness/Utility (appropriateness for the conference and symposium) Rate the manuscript on this criterion with a raw score from 1 to 5. A total calculated score will have a maximum value of five (5) and will be calculated from the four scored criteria using a weighting factor of 0.2, 0.3, 0.3, and 0.2, respectively. The weighing factor for this category is 0.2.): 5: Excellent 4: Superior 3: Satisfactory 2: Poor 1: Unacceptable Evaluation=Satisfactory (3) *** Raw Score for criterion (2) Technical Contribution (Raw Score for criterion (2) Technical Contribution (content, logic, development, etc.) Rate the manuscript on this criterion with a raw score from 1 to 5. A total calculated score will have a maximum value of five (5) and will be calculated from the four scored criteria using a weighting factor of 0.2, 0.3, 0.3, and 0.2, respectively. The weighing factor for this category is 0.3.): 5: Excellent 4: Superior 3: Satisfactory 2: Poor 1: Unacceptable Evaluation=Satisfactory (3) *** Raw Score for criterion (3) Quality of Content (Raw Score for criterion (3) Quality of Content (originality, novelty, completeness, references, etc.) Rate the manuscript on this criterion with a raw score from 1 to 5. A total calculated score will have a maximum value of five (5) and will be calculated from the four scored criteria using a weighting factor of 0.2, 0.3, 0.3, and 0.2, respectively. The weighing factor for this category is 0.3.): 5: Excellent 4: Superior 3: Satisfactory 2: Poor 1: Unacceptable Evaluation=Satisfactory (3) *** Raw Score for criterion (4) Quality of Presentation (Raw Score for criterion (4) Quality of Presentation (organization, grammar, language, graphics, etc.) Rate the manuscript on this criterion with a raw score from 1 to 5. A total calculated score will have a maximum value of five (5) and will be calculated from the four scored criteria using a weighting factor of 0.2, 0.3, 0.3, and 0.2, respectively. The weighing factor for this category is 0.2.): 5: Excellent 4: Superior 3: Satisfactory 2: Poor 1: Unacceptable Evaluation=Satisfactory (3) *** (1) Relevance/Timeliness/Utility (Appropriateness for the conference and symposium Comments on the manuscript related to this criterion): This is a work that study the response time of the requests in software system. A performance model is deveploed to estimate the response time. However, no Expirical validation is provided. *** (2) Technical Contribution (Content, logic, development, etc. Comments on the manuscript related to this criterion): In this paper the authors developed a SRN-based performance model for reactor pattern that provides synchronous demultiplexing and dispatching capabilities to network services and applications. The performance model can be used to obtain the response time of the requests that have two different processing priority: high and low. An interesting case study, the estimation of the response time of a VPN service provided by virtual router and its sensitivity to arrival rate, is involved as an application of this performance model. However, The proposed performance model is relatively simple. The authors may need to consider more cases, such as different source models except the Poission arrival and expentional service time. In addition, in the case study, more situations and deeper observations should be provided to enrich this paper. Swapna: I will add a small paragraph at the end which discusses how the SRN model can be extended to include non-Poisson arrivals, and also perhaps deterministic service times. *** (3) Quality of Content (Originality, novelty, completeness, references, etc. Comments on the manuscript related to this criterion): The accurate of the performance model is also not very clear. Expirical validation is still necessary to make this paper solid although the authors claim that this is their future work. Also, the conclusion section is rather terse. Swapna: I guess empirical validation is a whole different line of work, and we really cannot discuss that here. The novelty of this performance model is not very clear. it lacks stressing the difference with the relative work(e.g., [10]) and further applications of this model. Swapna: [10] uses SRNs for the event service, we are using it for another pattern. I will emphasize that in the paper. *** (4) Quality of Presentation (Organization, grammar, language, graphics, etc. Comments on the manuscript related to this criterion.): The flow of the presented work in this paper is good, which is roughtly similar to the reference [10]. The descriptions of both the model and the case study are clear. *** Other Comments (Please note any other specific revisions if any and give strengths, weaknesses, and supporting reasons for ratings.): In summary, I list three weakness: 1. The proposed performance model is relatively simple. 2. The novelty of this performance model is not very clear. 3. The accurate of the performance model is also not very clear. ===== Review ===== *** Raw Score for (1) Relevance/Timeliness/Utility (Raw score for criterion (1) Relevance/Timeliness/Utility (appropriateness for the conference and symposium) Rate the manuscript on this criterion with a raw score from 1 to 5. A total calculated score will have a maximum value of five (5) and will be calculated from the four scored criteria using a weighting factor of 0.2, 0.3, 0.3, and 0.2, respectively. The weighing factor for this category is 0.2.): 5: Excellent 4: Superior 3: Satisfactory 2: Poor 1: Unacceptable Evaluation=Superior (4) *** Raw Score for criterion (2) Technical Contribution (Raw Score for criterion (2) Technical Contribution (content, logic, development, etc.) Rate the manuscript on this criterion with a raw score from 1 to 5. A total calculated score will have a maximum value of five (5) and will be calculated from the four scored criteria using a weighting factor of 0.2, 0.3, 0.3, and 0.2, respectively. The weighing factor for this category is 0.3.): 5: Excellent 4: Superior 3: Satisfactory 2: Poor 1: Unacceptable Evaluation=Poor (2) *** Raw Score for criterion (3) Quality of Content (Raw Score for criterion (3) Quality of Content (originality, novelty, completeness, references, etc.) Rate the manuscript on this criterion with a raw score from 1 to 5. A total calculated score will have a maximum value of five (5) and will be calculated from the four scored criteria using a weighting factor of 0.2, 0.3, 0.3, and 0.2, respectively. The weighing factor for this category is 0.3.): 5: Excellent 4: Superior 3: Satisfactory 2: Poor 1: Unacceptable Evaluation=Poor (2) *** Raw Score for criterion (4) Quality of Presentation (Raw Score for criterion (4) Quality of Presentation (organization, grammar, language, graphics, etc.) Rate the manuscript on this criterion with a raw score from 1 to 5. A total calculated score will have a maximum value of five (5) and will be calculated from the four scored criteria using a weighting factor of 0.2, 0.3, 0.3, and 0.2, respectively. The weighing factor for this category is 0.2.): 5: Excellent 4: Superior 3: Satisfactory 2: Poor 1: Unacceptable Evaluation=Superior (4) *** (1) Relevance/Timeliness/Utility (Appropriateness for the conference and symposium Comments on the manuscript related to this criterion): The paper is about the performance analysis of a particular network function. This subject fits with the conference theme. *** (2) Technical Contribution (Content, logic, development, etc. Comments on the manuscript related to this criterion): An old modeling paradigm is applied for the performance analysis of a rather new networking solution. My major problem is that the elaboration level of the presented approach is very premature. The paper neglects obvious questions, such as - how relevant is the memory less exponential assumption on the arrival process and service time. - how accurate is the approximation applied in the paper. Swapna: I will add a small paragraph about the exponential assumption in the paper. *** (3) Quality of Content (Originality, novelty, completeness, references, etc. Comments on the manuscript related to this criterion): The presented system model is a very simple stochastic Petri net. Once the model is defined its analysis is obvious. Swapna: It is always the case, that in modeling the majority of the technical contribution is in the process of modeling. Once the model is done, the analysis is usually performed by tools. The definition of the model based on the real system behaviour is of some interest, but the paper does not present any considetration about the modelling process. Swapna: I do not quite understand this, also there isn't much room to explain the "modeling process" in the paper. The only visible point is that the author applied rude approximation at the more complex points. E.g., the number of type 1 customers arrives during the waiting time of a type 2 customer. *** (4) Quality of Presentation (Organization, grammar, language, graphics, etc. Comments on the manuscript related to this criterion.): Apart of some techincal detailes the presenation is clear and easy to follow. *** Other Comments (Please note any other specific revisions if any and give strengths, weaknesses, and supporting reasons for ratings.): page 2: the arrivals follow a Poisson distribution -> the arrivals follow a Poisson process Swapna: Fixed. page 2: Comment the validity of the exponential assumption !!! Swapna: A small paragraph has been added. table 1: - a bracket is missing - remove #Bi>=1 from the first two guard functions or add #StSnpSht=1 to the last two guard functions Swapna: I have added the missing bracket. However, I do not agree with the earlier suggestion, since I obtained these guard functions directly from the actual model. page 3: with weights given by trhe steady state probabilities -> with weights given by trhe steady state probabilities of the same system with on N1'=N1-1 (or N2'=N2-1) Swapna: I am not quite sure what N1' and N2' are in this case. I think there is something reasonable here, but I am just not able to understand what the reviewer is getting at. page 3: the time taken to complete -> the mean time taken to complete Swapna: fixed. page 3: unify n_1