The Wireless Mesh Networks (WMNs) is a wireless network contains two types of nodes such as mesh routers and mesh clients. In WMN, the mesh routers with minimal mobility form a backbone to mesh clients. Quality of Service (QoS) is a guaranteed level of service that improves the network performance and provides a guarantee for packet delivery. In WMN, the traffic prediction has a vital role in providing QoS guarantees to the communication. Compared to infrastructure less networks, the provision of QoS in WMN is simple due to the presence of minimal mobility mesh routers. The quality based routing protocol is used to optimize the delay and end to end probability of successful packet delivery in the wireless network, but it faces several routing challenges.

A wireless backbone is used for transmitting the data packets in a multi-hop fashion. Compare to MANET, the QoS provisioning in WMN is simple as it has fixed wireless mesh routers.

In WMN, the mesh clients equipped with a single interface that limits the network capacity. It is necessary to utilize the full available bandwidth to improve the QoS provisioning.

In WMN, channel access control and delay constraint are difficult due to mobile mesh clients. According to these characteristics, QoS provision is ambiguous in wireless networks.

The overhead data packets, route evaluation among nodes and mobility of nodes may affect the bandwidth in the wireless networks. The inaccurate bandwidth evaluation makes the communication inefficient.

Mesh routers equipped with multiple radio interface ranges and mesh clients equipped with a single radio interface. The mesh routers have high transmission power due to multiple radio interfaces, so the QoS provision is crucial in WMN.

         In WMN, the existing QoS routing protocols is classified into two types such as network layer and cross-layer. Figure 1 shows the classification of QoS routing in WMN.

Screenshot from 2017-10-09 14:26:29
Figure 1: Classification of QoS routing in WMN

         The network layer based QoS routing is classified into ad hoc based routing, controlled flooding routing, traffic aware routing and opportunistic routing.

         The ad hoc based protocols in WMN, admire with ad hoc routing protocols to contract with various link quality. The routers periodically maintain the routing information, and it is distributed among the nodes in the MANET. Many routing protocols used in ad hoc based WMN for example, Link Quality Source Routing (LQSR), and ScrRR.

         The QoS Routing Protocol (QoS-RP) [1] employs a broader path for data transmission in a hop-by-hop manner from source to destination. Each node evaluates a wider path to the destination and an in a trivial way the source node advertises all paths to the destination. Each node finds its own most extensive path makes routing overhead and the maintenance cost is high. To avoid the routing overhead this protocol employs isotonic path weight and selects a best widest path. The subpath of the widest path is allowed for transmission, only if the widest path fails to transmit. Each node in the network maintains two types of tables such as distance table and routing table. The distance table of the source node maintains the path advertises by its neighbors. A node evaluates the path itself is maintained in the routing table. If two adequate widest paths are available in the network, then the source selects one best path using a unique identifier. Suppose the two widest paths have same unique identifier, then the isotonic path weight comparison is used to find the best path. The hop-by-hop routing decision is used to selects the high capacity paths, and it satisfies the optimality and consistency requirements.

         In WMN, control flooding is proposed to decrease the routing cost. Compared with classical flooding, the controlled flooding is achieved in two ways such as temporal flooding and spatial flooding. In temporal flooding, the distance between two nodes is used to describe the frequency value. In spatial flooding, the distant node receives less comprehensive information from the source, so less flooding is enough to distant nodes. In a wireless network, a lot of connection occurs between two nearest nodes and the frequency of flooding is less between two nearby nodes. The path overhead is reduced due to the limited flooding. Many routing protocols are designed to control flooding, for example, Localized On-Demand Link State (LOLS) protocol, Mobile Mesh Routing Protocol (MMRP), and Optimized Link State Routing (OLSR).

         The Localized On-Demand Link State (LOLS) [2] is guaranteed to the packet transmission in the presence of several degraded links. It uses loop-free forwarding mechanism to find whether the link failure occurs in the network. In LOLS, a packet is transmitted in two modes such as greedy mode and recovery mode. The packet is forwarded to neighbors in greedy mode, and the destination path cost is decreased. In greedy mode, the packet strike at a dead-end. At that time the packet is forwarded in recovery mode. The packets in the recovery mode use a blacklist mechanism. Blacklist calculates the total number of failure links over a path. The routing information decides a path without including blacklisted links to next data transmission in recovery modes. After choosing an efficient path, the packet forwarding turns into the greedy mode, and blacklist links are clear. The LOLS routing protocol provides a loop-free path for communication even when the number of degraded links present in the network.

         In WMN, control flooding is proposed to decrease the routing cost. Compared with classical flooding, the controlled flooding is achieved in two ways such as temporal flooding and spatial flooding. In temporal flooding, the distance between two nodes is used to describe the frequency value. In spatial flooding, the distant node receives less comprehensive information from the source, so less flooding is enough to distant nodes. In a wireless network, a lot of connection occurs between two nearest nodes and the frequency of flooding is less between two nearby nodes. The path overhead is reduced due to the limited flooding. Many routing protocols are designed to control flooding, for example, Localized On-Demand Link State (LOLS) protocol, Mobile Mesh Routing Protocol (MMRP), and Optimized Link State Routing (OLSR).

         The Localized On-Demand Link State (LOLS) [2] is guaranteed to the packet transmission in the presence of several degraded links. It uses loop-free forwarding mechanism to find whether the link failure occurs in the network. In LOLS, a packet is transmitted in two modes such as greedy mode and recovery mode. The packet is forwarded to neighbors in greedy mode, and the destination path cost is decreased. In greedy mode, the packet strike at a dead-end. At that time the packet is forwarded in recovery mode. The packets in the recovery mode use a blacklist mechanism. Blacklist calculates the total number of failure links over a path. The routing information decides a path without including blacklisted links to next data transmission in recovery modes. After choosing an efficient path, the packet forwarding turns into the greedy mode, and blacklist links are clear. The LOLS routing protocol provides a loop-free path for communication even when the number of degraded links present in the network.

         Traffic-aware routing protocol uses common traffic matrix, and it follows a tree-structured network topology. The tree structure makes easy to maintain the routing information in the network. In WMN, the gateway requests to update the routing table for the frequent path evaluation. The gateway is working as a root of the tree. The addition and leaving of nodes is used to maintain the route in WMN. The Multipath Routing Algorithm based on Traffic Prediction (MRATP) [3] is used to improve load balancing scheme and keeps a maximum number of the disjoining path for efficient communication. The non-linear traffic prediction technique reduces the noise measurements in the network. The non-linear traffic prediction model contains wavelet and neural network models. The wavelet model accurately predicts the network traffic without any real-time characteristics. This protocol combines both wavelet and neural models to achieve high traffic prediction. It is explained in three types of aspects. They are multipath route discovery, traffic prediction due to congestion detection, and load balancing techniques using multipath routing.

         The opportunistic routing protocol supports the successful retransmission of data packets at the time of link failure between two communicating nodes. Opportunistic routing protocol provides at least one guaranteed hop for successful packet delivery. The high delay occurs because it takes more time to find the failure links in the network. Several types of routing protocols are used in opportunistic routing, for example, Extremely Opportunistic Routing (ExOR) protocol and Resilient Opportunistic Mesh Routing Protocol (ROMER).

         The designation of ExOR in [4] protocol faces four challenges. The nodes should agree on the receiving packet identities, and it selects one forwarder using the identity. The agreed protocol has the lowest overhead and is capable of forwarding the packet to the destination at many times. The agreement cost increases due to the number of nodes. In ExOR protocol, the sender disseminates a data packet to its neighbors. The data packet employs batch map contains the highest priority node that is closest to the destination approximately to transmit all packets. The particular nodes not only receive the packet either group of intermediate nodes are receiving the same data packet. If the highest priority node disseminate one data packet, and it receives an acknowledgment for the sending packet. If the acknowledgment is not received, then the batch map chooses another high priority node for retransmission. It improves the throughput between two distant nodes, and it reduces the unnecessary retransmissions using acknowledgment. The ExOR improves the network capacity and also increases the throughput gain. However, the duplicate transmission occurs when wrong coordination among nodes.

  • Hou, Ronghui, et al. “Hop-by-hop routing in wireless mesh networks with bandwidth guarantees.” IEEE Transactions on Mobile Computing,2012.
  • Robertson, G., & Nelakuditi, S. “Handling multiple failures in IP networks through localized on-demand link state routing”’ IEEE transactions on network and service management,2012.
  • Zhi-yuan, Li, Wang Ru-chuan, and Bi Jun-lei. “A multipath routing algorithm based on traffic prediction in wireless mesh networks.” Fifth International Conference on Natural Computation, ICNC’09,2009.
  • Biswas, S., & Morris, R. “ExOR: opportunistic multi-hop routing for wireless networks”, ACM SIGCOMM Computer Communication Review,2005.
Leave Comment

Your email address will not be published. Required fields are marked *

clear formSubmit