We have witnessed in the last decade a significant increase in the use of multiple wireless networks within IoT architectures which are heterogeneous in terms of available resources, downlink data rates and buffer sizes. In addition to that, the IoT flows are characterized by different requirements such as the tolerated latency and data rates. Therefore, it is crucial to design mechanisms that dynamically and optimally selects the network interface that satisfies the flows’ requirements and maximize the accepted flows by an IoT gateway.
In this thesis, we model the problem of assigning flows in a multi-interface IoT gateway. We define the flow optimization problem as a non linear integer programming (NILP) which maximizes the amount of accepted data by the gateway while satisfying the available network resource capacity and the requirements of the IoT flows. In addition, we design a handover protocol to migrate flows from one interface to another and a control module within the gateway that takes the decision of the flow assignment. Towards this, we propose two algorithms to solve the optimized flow assignment problem (OFAP). The first solution is based on the greedy approach and the second is based on dynamic programming. These two algorithms provide optimal assignment of flows in a multi-interface IoT gateway.
Furhermore, we design and implement the control module and the handover protocol and we provide simulation results that show the effectiveness of our algorithms.
| Date | 14 Aug 2019 |
|---|
| Original language | French |
|---|
| Awarding Institution | - École de technologie supérieure
|
|---|
| Supervisor | Mohamed Cheriet (Supervisor) & Kim Khoa Nguyen (Co-supervisor) |
|---|
Amor, M. G. (Author),
Cheriet (Supervisor) &
Nguyen (Co-supervisor),
14 Aug 2019Student thesis: Master's thesis › Master in Engineering: Information Technology Engineering