A Simulation Performance Study of PSN Algorithms in Practical Use Cases
dc.contributor.advisor | Makaroff, Dwight | |
dc.contributor.advisor | Stanley, Kevin | |
dc.contributor.committeeMember | Eager, Derek | |
dc.contributor.committeeMember | Berscheid, Brian | |
dc.contributor.committeeMember | Keil, Mark | |
dc.creator | Zare Abandankeshi, Fatemeh 1991- | |
dc.creator.orcid | 0000-0001-9098-8187 | |
dc.date.accessioned | 2020-02-21T14:44:26Z | |
dc.date.available | 2020-02-21T14:44:26Z | |
dc.date.created | 2020-01 | |
dc.date.issued | 2020-02-21 | |
dc.date.submitted | January 2020 | |
dc.date.updated | 2020-02-21T14:44:27Z | |
dc.description.abstract | Delay Tolerant Networks (DTNs) could have substantial value in areas where Internet infrastructure is expensive or dangerous to deploy. A DTN consists of a set of nodes that can transfer messages to each other and immediate packet delivery is not necessary. Pocket Switched Networks (PSNs) are a special case of DTNs where packets are forwarded based on the historical contact patterns between nodes which are assumed to be mobile agents like people or animals. Routing is a challenge in PSNs since an end-to-end path is unlikely to be available from source to the destination. Previous works proposed the idea of utilizing the social behavior of human contacts to apply different decisions for routing based on social clustering. These ideas can improve PSNs performance in terms of delivery ratio, energy consumption, and delivery delay because transmitting messages around a group is easier due to the higher probability of contact between source and destination. Contact stability and diversity, network resource capacity, clustering algorithms, and the transmission range of devices may affect the performance of PSN routing algorithms. In this thesis, the effect of each of these parameters on the performance of PSNs algorithms is evaluated by different use case scenarios. Evaluating the performance of PSN routing algorithms with different circumstances requires a framework that supports cluster-based routing algorithms. Previous DTN simulators do not explicitly support cluster-based routing algorithms. In this thesis, a DTN simulator, PYDTNSIM, has been extended to compare different available cluster-based routing algorithms. This simulator is modular and can be extended for the implementation of other routing and clustering algorithms. Currently, it supports three different clustering algorithms and three routing algorithms (two are cluster-based, and one is unclustered). PYDTNSIM can compare the performance of different PSN routing algorithms in terms of delivery delay, delivery ratio, number of message copies generated, and energy consumption. The simulator can track message transmissions in intermediate nodes, as well as variable message size and several message generation heuristics. To evaluate the effects of clustering techniques on the performance of PSN routing algorithms, several clustering algorithms are deployed to cluster network nodes. Advanced Graph-based Kmeans (AGKmeans) is proposed in this thesis as a clustering algorithm by using the Kmeans clustering concept. This algorithm is appropriate for datasets in which the participants are dynamic and the dataset can be modeled as a graph. The initial centroids selection in AGkmeans is not performed randomly. To evaluate PYDTNSIM with different experimental parameters data analysis of the realistic datasets, SHED1 and SHED5, and virtual dataset generation emulator, the Termite, is done to extract the contacts from different environments with different transmission ranges. | |
dc.format.mimetype | application/pdf | |
dc.identifier.uri | http://hdl.handle.net/10388/12675 | |
dc.subject | delay-tolerant networks | |
dc.subject | social-based routing | |
dc.subject | resource-constrained devices | |
dc.subject | clustering | |
dc.title | A Simulation Performance Study of PSN Algorithms in Practical Use Cases | |
dc.type | Thesis | |
dc.type.material | text | |
thesis.degree.department | Computer Science | |
thesis.degree.discipline | Computer Science | |
thesis.degree.grantor | University of Saskatchewan | |
thesis.degree.level | Masters | |
thesis.degree.name | Master of Science (M.Sc.) |