Modeling and Performance Evaluation

To solve the problem of vertical handover in heterogeneous wireless networks, this paper proposes a seamless handover method for heterogeneous wireless networks based on software defined network (SDN) architecture. The proposed method requires SDN control plane to periodically collect attributes of wireless networks so as to evaluate network performance. Considering varied types of services have different requirements and users have different preferences, fuzzy analytic hierarchy process (FAHP) and multi-path transmission control protocol (MPTCP) are applied to select the most appropriate target network and execute seamless handover, respectively. Compared to existing methods, the proposed method jointly considers service requirements, user preferences, and mobility management for network selection and guarantees service continuity during vertical handover. Simulation results show that the proposed method makes a comprehensive decision on network selection. In terms of handover times, the proposed method achieves up to 17.4% and 4.5% reduction compared to received signal strength based algorithm and fuzzy multiple criteria group decision making based algorithm, respectively. The proposed method also maintains at least one subflow connected during the handover process and achieves seamless handover.

A FAHP and MPTCP Based Seamless Handover Method in Heterogeneous SDN Wireless Networks

Increasing number of connected devices and the volume of data generated by these devices pushing current generation of cellular network to their limits. Cross layer designs (CLD) are recommended for the designing a better performing network applications. 5G standard for next generation network includes the provisions for CLD within the standard itself to cater resource demand from such devices. 5G along with its predecessors is going to coexist providing a heterogeneous connectivity for a user. Irrespective of generation, radio access network (RAN) still remained a bottleneck and drastically affects the performance. Signal strength is a metric for RAN which is used in several decisions and is an important parameter for CLD. In simulation of CLD signal strength is modeled using established path loss and fading models wherein distance is the only variable, similarly, noise is modeled as a Gaussian distribution. Thus, with static user received signal strength (RSS) appears as Gaussian distribution but in reality the RSS is non-uniform with sharp peaks and long tail. Simulation with incorrect models result into solutions that lag far apart in performance than the actual outcomes, hence, correct modeling of RSS is a decisive factor. In this empirical study data collected for RSS using NeSen App by colocated multiple smartphones. This study proves that RSS recorded follows t-Distribution with mean between [-71, -77]dbm, standard deviation between [3.75, 4.85] and degree of freedom in range of [6.8, 10.5].

R K Mishra

Papers

RSS Characterization in Wireless Heterogeneous Network