1. Introduction

The 802.11 LAN [3] have been adopted by various industries such as education, office, home etc… is not able to satisfy the current demand of end user need due to the transmission power constrained imposed by the regulatory body which limit the coverage of network. Since these networks are wired in nature they are inflexible and induce high deployment cost, to address this task group S was formed and wireless mesh network is standardized as 802.11s [2]. In next generation the WMN [19, 20] is anticipated to be the most sorted out paradigm for providing internet connectivity to its end user by complementing existing wired network. WMNs establishes an ad-hoc infrastructure i.e. it is self-configured and self-organized, with the devices in the network can join and leave a network and still maintaining the mesh connectivity. In existing wireless network such as Wi-Fi and 3G network are connected to AP (Access Point) through wired infrastructure in order to access/ retrieve the information from the backhaul infrastructure which is time consuming and induce high cost of deployment. Subsequently WMN uses subgroup of access points to communicate with internet gateway (IGWs) through wired infrastructure, whereas the other access points, namely knowns as mesh routers (MRs) are connected through internet gateway in a multihop based communication. Thus WMN help in providing an economically feasible and sustainable solution for wireless internet connectivity.

To provide quality application services to wireless mesh devices by adopting an IEEE 802.11s mesh protocol standard is known mesh stations. The mesh station that offers admission to the distributed network through the wireless medium is known as mesh gate. The other Mesh stations work as MAP (Wireless Mesh Access Point) to its end devices which are mobile in nature. The MAP acts as coordinator node to the gateways as well as delivers data to its end user and they are equipped with multi radio to provide multi-channel access for robust communication. The architecture of IEEE 802.11s wireless mesh network is shown in Fig. 1.

Figure 1. The IEEE 802.11s wireless mesh network architecture

There exist a load balancing issues in wireless mesh network backbone mesh point that act as a gateway to connect other mesh point which are in it routing path or to internet. Due this it is necessary to balance the load considering dynamic nature of traffic and speed of mobile user and also the nature of wireless medium transmission channel. The selection of gateway router by the mesh point is crucial since all traffic of end devices request will go through the gateway so there is high likelihood of bandwidth bottleneck resulting buffer overflow, packet loss which is result in congestion in network, so load balancing is an important requirement to improve the network performance.

There has been various protocols have been adopted to reduce network overhead on wireless link by distributing traffic [4] [5].

In [11] they showed hybrid wireless mesh protocol is not efficient enough for random access MAC, and to overcome this they devolved a flexible model that properties of random access for hybrid wireless mesh protocol. In [12] they proposed a prototype namely HWMP+ which is an enhancement of [11] which consider the channel condition and load in network for selection of root. Their method improved the network performance but they did not considered the multi radio channel allocation scenario. In order to address the channel allocation issues in wireless mesh networks. In [13], [14], [15] they considered fixed channel width adaptation for a single-radio multichannel, their outcome are not suitable to the OFDMA. In [16] [17] they extended the protocol by considering multiradio multichannel and this is also not suitable for OFDMA due to issues of each device can either transmit or receive data on sub channels at a slot time, while this issue is not bid factor for multiradio multichannel circumstances. In [18] they adopted energy dissipation and improve the network performance author considered the power against network data rate by the tradeoff curve. Physical interference involved where device used for controlling the power. Author work is based on single channel or single device; it is not suitable for multi-channel or multiple user scenarios.

The fixed data rate contention based CSMA [6] is not efficient in providing high throughput services in order to achieve this OFDMA or TDMA is adopted. The issue with OFDMA [7] [9] based protocol is that it has transmitting and receiving in a slotted time, and sub channel allocation of radio can either be used by receiver or transmitter. Channel width adaptation [8] consider adaptation of time and frequency domain. To overcome this TDMA is been used. The issue with TDMA based approach is that improper measure of active user will result in improper usage of frequency as a result throughput performance is affected to overcome this spatial reuse based TDMA [10] is adopted but it was not suitable for multi-channel and multi user scenario due to interference of user. To overcome these challenges here the author propose threshold selection based decision making algorithm for efficient mobility management by adopting a cross layer design considering link quality and evaluate the performance of proposed approach with existing work DGR-SRORA [9].

The organization of the work is as follow: In section two the literature survey is briefed. In section three the list of symbol and notation use is tabulated. In section four the proposed threshold selection algorithm is presented. In section five the experimental study and evaluation is done. In penultimate section the concluding and future remark is discussed.

2. Literature Survey

Here the author surveys some of the existing protocol that tries to improve the QoS issue of WMN. The intrinsic issues of WMNs, such as power limitation, high BER and interference, considerably limit the WMNs performance. In the past many studies and methodologies [21-26] have been presented by researcher to enhance the time slots for channel allocation and throughput in channel utilization of WMNs. Nevertheless, these methodologies do not consider collision detection mechanism considering the mobility nature of user in mesh network.

Various methodologies have been proposed in recent times for performance enhancement of IEEE 802.11s wireless mesh network that adopt CSMA/CA MAC protocol [27] [28] that include tuning the carrier-sense range [29] by adopting out of band control message [30] or local coordination [31] in order to improve the channel slot utilization. The CSMA/CA protocol used in IEEE 802.11 [32], are compromised by the issues of application service unfairness and inefficiency to address the packet delays. Though these issues have been addressed in [33] the Bidirectional-DCF (BDCF) protocol, on IEEE 802.11 DCF protocol, the issue still persists in multi-hop wireless mesh networks [34], [35].

The TDMA based MAC protocols [36] [37], offers collision-free transmission and provide good access control of throughput and delay of network traffic. With the institution of TDMA-based scheduling the IEEE 802.11s [38] was drafted for mesh network. By using the mesh coordination function (MCF) in IEEE 802.11s, each link can reserve a slot time for data transmission without competing with other links for transmission opportunities. However, the end-to-end QoS for delay sensitive services is still an open issue. To address the resource-allocation issue of multihop OFDMA for wireless mesh networks has been developed in [39-42]. Nevertheless, these methodologies focused on a relay-based two-hop network prototype which is not suitable for mobility nature of mobile user in WMN. The MAC protocols based on distributed scheduling [43], [45], [46] have been proposed to address the interference free [44] TDMA schedules. Yet these methodologies allocate equal slot access on time basis for each user request irrespective of traffic demand or load which affect the end to end performance. Along with this there has been various cross layer design have been adopted to solve the routing, interference and MAC contention for better path selection [47] but they suffer in identifying alternative path and failed to address the interference problem.

It is seen from literature survey there is need for new protocol design that consider load among mesh point, mesh gateway, link quality and mobility of user end device. To address this here the author proposes a threshold based mobility management protocol. The ideal threshold should satisfy maximizing the throughput and minimize the packet loss/drop which is briefed in section three.

3. Proposed Threshold Based Mobility Management Model

This work proposes a cross layer threshold based decision making strategy for mobility management by considering the link quality and threshold detection strategy based on collision detection algorithm to detect the set of active user in wireless mesh point. Here the author consider a wireless mesh network that has a set of devices that associates with wireless mesh access points and . Here the author consider a multi radio TDMA based channel and also consider infinite buffer size for obtaining packet drop due to failed or wrong transmission. To optimize the data transmission in wireless mesh network for packet collision avoidance here the authors consider an OTS (optimization transmission stage) as in [1] and the threshold selection algorithm for efficient decision making strategy for mobility management is presented in below subsection.

3.1. Threshold Selection Algorithm for Mobility Management

Here the author considers a case were every wrong judgment taken by the wireless mesh access point will outcome in failure of packets in OTS so an effective threshold selection algorithm is needed to detect the set of active user in the network for efficient mobility management for wireless mesh networks. Once the mesh point detects the collision correctly it helps in improving the throughput of the network. So it is important to find the best threshold selection to assure a promising max throughput.

(1)

where H is the Promising Throughput, is the likelihood of Rayleigh flat fading channel as in [1].

(2)

Where is likelihood of buffer being empty for user at initialization of OTS, is the likelihood of detection under Raleigh flat fading channel as in [1], , , .

The equation (2) can be simplified by considering peak N signal to noise ratio (SNR) by considering and .

(3)

(4)

(5)

From the equation (4) & (5) the author computes the likelihood of ideal threshold on Rayleigh flat fating channel as follows

(6)

The Ideal threshold for efficient mobility management is stated as follows.

(7)

Based on equation (7) the maximum throughput can be obtained for the mesh point but to obtain ideal threshold detection for realistic mesh environment the author consider a scenario were few packet that are collided can be recovered by the mesh access point makes a false decision. I.e. when sources are lost in detection, out of that at least packets that are collided can be recovered correctly in the OTS. The mean probable packet recoverable is as follows

(8)

The probable throughput under this scenario is

(9)

Considering when the author expect that and , therefore .

To obtain ideal threshold detection under this scenario the author considers that let Based on above Eqs. (8), (9) and also Eq. (21) from [1] the author obtain

(10)

When considering good N, the Eq. (10) is simplified when is approximated is as follows

(11)

Now let consider Eq. (12) form [1] the likelihood packet loss is formulated as follows

(12)

The function can changed to likelihood packet loss . i.e. let . Therefore based on Eq. (12) we obtain

(13)

When considering good N, the Eq. (13) is simplified when is approximated is as follows

(14)

Based on this the ideal threshold can be improved from Eq. (7) and based on this the selection of mesh point for user is considered.

Beyond this threshold the packet loss increases and the throughput is dropped. This indicates that the user is moving away from the Mesh Point. Below in the next section the experimental analysis study of our proposed approach is evaluated and compared with existing algorithm namely DGR-SRORA [9].

Table 1. List of Symbol and Notation Used

4. Experimental Analysis

The operating system used is 64-bit windows 10 single language, 64-bit quad core processor 64-bit with 12GB of RAM. We have used MATLAB 2015 framework. We have conducted experimental analysis on following parameter for drop rate for varied user speed and throughput for varied user and speed and compared our proposed approach with existing algorithm DGR-SRORA [9] and conducted experimental analysis.

In Fig. 2, the user speed is varied from 5 to 40 m/s and the throughput performance is shown. The throughput performance improvement of proposed approach over existing DGR-SRORA when speed is 5 m/s is 9.76%, for 10 m/s is 21.20%, for 20 m/s is 20.11% and for 40 m/s is 36.38%. It is seen from the graph the throughput achieved is decreasing with increasing in speed for both the approaches which proves Eq. (11) and the average throughput improvement of proposed approach is 17.83% over existing DGR-SRORA which is shown in Fig. 3.

The experimental parameter considered is shown in below Table 2

Figure 2. Throughput performance for varied speed

Figure 3. Average throughput performance for varied speed

In Fig. 4, the user device is varied from 5 to 20 and the throughput performance is shown. The throughput performance improvement of proposed approach over existing DGR-SRORA when user is 5 is 33.33%, for 10 is 47.15%, for 15 is 32.21% and for 20 is 52.07%. It is seen from the graph the throughput achieved is decreasing with increasing in number of user in the mesh network for both the approaches which proves Eq. (9) and the average throughput improvement of proposed approach is 46.13% over existing DGR-SRORA which is shown in Fig. 5.

Figure 4. Throughput performance for varied user

Figure 5. Throughput performance for varied user

In Fig. 6 below, the user speed is varied from 5 to 40 m/s and the drop rate performance is shown. The drop rate performance improvement of proposed approach over existing DGR-SRORA when speed is 5 m/s is 30.88%, for 10 m/s is 40.05%, for 20 m/s is 37.02% and for 40 m/s is 33.78%. It is seen from the graph the drop rate achieved is increasing with increasing in speed for both the approaches and the average drop rate proposed approach is reduced by 35.09% over existing DGR-SRORA.

Figure 6. Drop rate performance varied speed

In Fig. 7 below, the user speed is varied from 5 to 40 m/s and the packet processing delay performance is shown. The packet processing delay is reduced by proposed approach over existing DGR-SRORA when speed is 5 m/s is 33.3%, for 10 m/s is 33.41%, for 20 m/s is 17.32% and for 40 m/s is 14.78%. It is seen from the graph the packet processing achieved is increasing with increasing in speed for both the approaches and the average packet processing delay of proposed approach is reduced by 24.16% over existing DGR-SRORA.

Figure 7. Packet processing delay performance for varied speed

5. Conclusions

Here the author presented an efficient mobility management scheme for WMNs that meets the QoS necessities of the end devices. Here in this work the author have presented threshold based decision making algorithm for efficient mobility management in mesh networks by adopting a cross layer design. The proposed model minimizes the packet drop, reduces the packet processing delay and improves the throughput performance in WMNs due to the adopted threshold selection technique. The experimental result shows that the proposed model reduces drop rate by 35.09% for varied speed, reduces the packet processing delay by 24.16% and improves the throughput performance by 17.83% for varied speed and 46.13% for varied user over existing DGR-SRORA. The outcome shows that the proposed model is robust and scalable considering network density and mobility of user. In future the author considers evaluating the performance of proposed model for varied load density.