Conclusions and Future Scope 1 Chapter 7 Conclusions and Future Scope 7.1 Summary of Experimental Finding The value of network simulator in the communication field is indispensible as they provide support for real-time models with customising possibility, while at the same time being relatively cost effective. It also helps in analysing the Search and Rescue operations in post Disaster Management. To conclude, we have compared NS3 and OMNET++ simulators that involved the MANET routing module with respect to different network parameters.we compared MANET Routing protocols with different node scenarios i.e. Worst scenario (15 nodes), Average scenario (25 nodes) and Best scenario (50 nodes). Finally, a comparison has been made both on OMNET++ and NS3 results based on selected parameters, i.e.; packet delivery ratio, average end-to-end delay, throughput, routing overhead and path optimality. All these experiments were conducted on a machine with an Intel core i72.4ghz 8 GB RAM and Ubuntu 14.04(64 bit) and NS3 version 3.23 and OMNET++ version 4.4 and INET 2.3.0 framework were used for network simulation. Each experiment was run for a simulation time of 200seconds. From our observations we can say that NS3 is better with some parameter and OMNET++ is also good in selective network parameters. It can also be analysed from the results that NS3 is better when consistency is taken into consideration while OMNET++ have significant amount of variation in results. When the source codes of different routing protocols are compared on simulators, it was observed that they both follow the same version with few protocols. For example, the source code of AODV protocols was analysed in particular, the outcome showed that both the simulators had same versions of AODV protocols implemented in it. Hence, it can be said that the difference in the performance of these simulators is not due to different source codes; it is due just to the manner in which these simulators carry out the simulation. To validate the
Conclusions and Future Scope 2 accuracy of network simulators, we compare the network simulator with real-time simulation. We checked the accuracy of network simulator and real-time simulation with average end to end delay, overhead and Throughput with 15 nodes in the same simulation setup. Experimental results show that the NS3 simulators along with MANET reactive model are able to work very well in real time simulation for 15 nodes. While OMNET++ simulator does better with proactive protocols for 15 nodes. Through this study we are able to show that it is possible to improve the execution speed by varying simulation time and event procedure. To achieve this objective a separate parameter int 64x64_t in a fixed point integral had to be included to support a 64 bit integer part and a 64 bit fractional part where, x denotes the position of fractional point. Using conversion operator we can convert time to new integral type. This increased the execution time of the simulator. For the enhancement of event procedure, we follow the hybrid approach. For sustaining maintainability in code, use switch case and for proper readability, creates an event handler. It is possible to keep the switch based code organization in simple protocol types, but when protocol complexities increase, it is easy to shift the event handler with different arguments to be kept apart and results of this presents that these slight changes improve readability, maintainability of code of program and its also improve the efficiency of simulator performance. As a result, we may conclude that evaluation of protocols from different simulators is not comparable even when we use the very same simulation scenario and source code. Based on the experience of research, it can be proposed that modules, i.e. the implementation of a particular model or protocols, should be the level of abstraction on which different model and protocols implementations should be compared. Also at the same time, the analysis and evaluation were done through network connectivity as we have considered different number of nodes for simulations. Hence,we can say that in future the same study can be done by varying network simulation time, bandwidth or by varying mobility pattern with same or even other network for comparison of results with NS3. 3G,4G, LTE and WiMAX technologies which are the latest, the performance metrics can be taken into account for advanced communication. 7.2 Contribution and Accomplishment of Research The answers to the questions proposed in this study are as follows:
Conclusions and Future Scope 3 General Objective: A unified analysis of Reactive, Proactive and Hybrid Routing Protocols for Mobile Ad Hoc Network in order to check feature based performance of Simulators Accomplishment:- from the experimental detail study we conclude the following:- We have selected Proactive (DSDV & OLSR), Reactive (AODV & DSR) and Hybrid (ZRP & WARP) protocols from the table no. 1, 2 and 3 in chapter 2. For experiment purpose, selection criteria of simulators mentioned in table 5 in chapter 2. These are NS3 and OMNET++ popular event based simulators selected, that offers a wide variety of models to use. AODV reactive protocols performance with each selected parameter better than other proactive and hybrid protocols. Performances of protocols are better with average and best node scenarios are in NS3 simulator. Specific Objective: In order to achieve the general objective following specific objectives are drawn: 1. Review the significant challenges in popularly used network Simulators that hinders their performance Accomplishment:- CPU Utilization, Memory Consumption, Scalability, Reliability, Troubleshoot for different network models and protocols, Programming Languages, Documentation and Available supports for different network components. 2. A unified analysis of Reactive, Proactive and Hybrid Routing Protocols for Mobile Ad Hoc Network in order to check feature based performance of Simulators Accomplishment:-for this question we conducted the experiment of Reactive (AODV & DSR),Proactive ( DSDV & OLSR) and Hybrid (ZRP & WARP) protocolson NS3 and OMNET++ based on network parameters PDR, Throughput, Average End-To-End Delay, Routing Overhead and Path Optimality. These all results are highlighted in chapter 5. Results in both network simulators are different. 3. Find out the suitable simulation features and how they affect the simulation results
Conclusions and Future Scope 4 Accomplishment:-features of simulators which affects the simulations results are API, code size of a programme, real world module(real byte), fragmentation, reassembly, memory efficiency,pretty printing, simulation data and extensibility. 4. Select specific feature that can play a significant role in improving the overall performance of simulator in real world scenarios Accomplishment:-here we selected only two features of simulators for improvements. They are Time and Event procedure. 5. Check the credibility of Simulated vs. real world results in order to validate the results of Simulators with Routing Protocols for Mobile Ad Hoc Network Accomplishment:-For this we conducted an experiment with 15 nodes (12 smart phones and 3 laptops)by created ad hoc network with them and evaluated results based onthroughput, Average End-To-End Delay and Route Overhead. Details mentioned in chapter 5. 7.3 Future Work In this age, research work on network is complicated and this is due to the fact that it complicated update process that takes place on every mint. Despite all this there are always new methods; new implementation techniques and new protocols that are availablein the market. None of the existing papers ever compared any routing protocols with network simulator. This thesis was exceptional in that this study presented a comparative study and results of NS3 and OMNET++ open source simulator for analyzing Mobile Ad hoc routing protocols by means of Average End-to-End Delay, Packet Delivery Ratio, Path Optimality, Routing Overhead and Throughput. By means of this research was analyzed which one of the open source simulator is best suited for any research work. In this research we have presented proactive, reactive and hybrid routing protocols which are all compared with both network simulators. Experimental result exhibits that NS3 shows higher values for all the network parameters (Average End-to-End Delay, Packet Delivery Ratio, Path Optimality, Routing Overhead and Throughput) when compared with that of OMNET++ for different MANET Protocols. Real world experiment also shows that with respect to reactive protocols, NS3 performs better than OMNET++.
Conclusions and Future Scope 5 Considering all the above, it is inferred that NS3 is a network simulator for future scholarly investigations, it will completely supplant some other simulator because of its elite accessible elements, system models and documentation. Then again, OMNET++ is driving business commercial enterprises because of better utilize backing and documentation. The same analysis can be simulated with OMNET++ or done other network simulators for comparing the relative results. In the future, execution assessment measurements can be executed for cutting-edge correspondence advances, for example, LTE, VoIP, 3G, 4Getc. One enhancement would be to consolidate the accessible Discrete Event and the Continuous reproduction strategies to shape a "50:50 Simulator or "Hybrid Simulator", to execute the performance evaluation measures and watch the code of conduct of different network simulators.