This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
In recent year, the Mobile Ad-hoc Networks (MANETs) have become more and more popular and useful, and lots of research communities also have concentrated with deploying. Due to the MANETs provide a flexible way to build up communication circumstance with geographical constraints such as disaster (earthquake), battlefield environment and others . Moreover, in the communication process, the routing is playing an important role. The task of routing is forwarding packets from a source node to destination node. But this task is difficult to implement in MANET networks. Because the MANET networks are mobility every node can serve as a router and lack of infrastructure. These factors promote the routing algorithms must be robust and adaptive in a self-configuring and decentralized network . This document will investigate some challenges that are faced by routing in MANETs and incorporation intelligence routing (router algorithms) to improve the challenges.
2. Challenges in MANETs routing
The MANET network is often lack of central control and each node (acts as a router) can free to move randomly. And the topology changes rapidly and unpredictably in MANET network .These unique features are nontrivial challenges of MANET routing. And these characteristics are leading to the consequence that routing is more difficult to apply in the MANET than wired network . In order to achieve these challenges the developer need to incorporate the intelligence routing. This part will discuss in next section. In this section the document claims some challenges that are faced by routing in MANET network.
2.1 Lack of Infrastructure
The MANET network is a decentralized mobile wireless network. It promotes each node can forward packets for other nodes, and the determination of which nodes forward packets are made dynamically based on the network connectivity . In other word, each node can act as a router, and they can forward packets to any other nodes and multiple paths to transfer. Therefore the traditional wired routing algorithm cannot support in MANET network such as the distance-vector routing (RIP) and link-state routing (OSPF). Due to the traditional routing protocols need cooperative all nodes and the infrastructure and unidirectional links between nodes .
2.2 Mobility of the MANETs
The second challenge in the MANET routing is mobility (scalability) problem. The scalability is predefined when it is designed and not changes during the use in the traditional wired network . In contrast, the scale of the MANET networks is always changed. Due to each node acts a router leading to mobility in the MANET networks. Moreover, the nodes maybe disappear and reappear because of the mobility . Thus the developer difficult predicts how many nodes in the network. As a result, the routing protocols are generating more errors than wired network and difficult to implement in MANET networks. It requires that routing protocols can be compatible with the changing scale of the MANET networks.
3. Intelligence in MANETs routing protocols
Intelligence in MANET routing protocols, it means that address good routing algorithm to achieve good performance including find the optimal path, minimise latency, minimise packet loss and reduce failure in transmission and so on . In traditional wired network, these features can easy achieve to use the wired routing protocols such as RIP and OSPF. However, these performances are difficult to achieve in MANET networks. Because of the MANET network is mobility and lack of central control. So it requires the routing algorithms must be robust and adaptive in a self-configuring and decentralized network . And traditional wired network routing is hard to support in MANET network. This document will focus on intelligence part to investigate and discuss how they can help MANET network to achieve good performance. There are three types routing algorithm supported in MANET network included table-driven algorithm (proactive), demand-driven algorithm (reactive) and hybrid algorithm . The paper will investigate these algorithms which can improve performance for MANET. A [Figure 1] shows MANET routing protocols as following:
3.1 Table-driven algorithms
Previous part the paper has addressed the task of routing is forwarding packets from a source node to destination node. And the procedure of forwarding packet is very simple which dependent on the routing table. The transfer nodes find the destination address from the routing table then forward it. However, in the MANET network build up the routing table in the nodes are very difficult. On the other hand, the table-driven algorithm also means proactive algorithms, such as DSDV , OLSR  and FSR , can solve this problem. It supports all nodes maintain routing tables all the time.
In traditional routing protocols, the RIP and OSPF are also proactive routing protocols. But they cannot support in MANET network. Firstly, these protocols need consumes lots of bandwidth when they are implementing in network. The wired network can deploy several hundred gigabit bandwidth while the bandwidth of 802.11b is only 11 megabit. As a result, the routing communication is overhead. The Destination-Sequenced Distance Vector Routing Protocol (DSDV) is abandoned due to its high communication overhead . Secondly, the nodes in MANET network can free to move randomly. It leads to the topology changes rapidly and unpredictably . Thus the MANET network requires proactive routing protocols updated information of each node (routing table) more frequently than wired network.
The Optimized Link State Routing Protocol (OLSR) is based on the link-state algorithm . It supports the Multipoint Relaying (MPR) flooding technique  and Dijkstra's shortest-path algorithm to solve these problems and achieve good performance in network communication. It can be seen as an intelligence part of OLSR algorithms in MANET network. Traditional, all nodes flood neighbour information in a link-state protocol, but not in OLSR. Because of it send information into its multipoint relay selector set (MPR set).The MPR technique can reduce flooding node information, reduce the number of topology broadcast packets and reduce the number of retransmissions packet. Therefore lots of bandwidth is saved in communication process. An example is showing how MPR technique runs in the OLSR routing protocols as following:
At the first, the [Figure 2] shows that all nodes periodically broadcast 'HELLO message' to their one-hop neighbors from node 4 and the 'HELLO message are not forwarded'. In other ward, the node 4 wants to communication with another node, the node 4 will broadcast a route request packet to its neighbors.
Secondly, the node 4 receives the neighbor list, it helps to determine their two-hop neighborhood and an optimal MPR set [Figure 4].
The result of this example is transmission data to node 3 and node 6. All nodes can be received the data in the network. So the MRP technique can reduce duplicate transmissions and reduce size of control packet. It helps MANET routing protocol save much more bandwidth when they are implementing in network. And the Dijkstra's algorithm support in OLSR to find shortest path in network transmission. In addition, it uses a broadcast route discovery mechanism and hop-by-hop routing. These features reduce response time to routing requests and trigger updates when necessary.
3.2 Demand-driven algorithms
The MANET network is mobility leading to the topology changes quickly and unpredictably. Therefore each node must store and update information frequency in network communication, and it spends lots of bandwidth. However, sometimes this spend is a waste of bandwidth . In order to reduce the waste of bandwidth, the demand-driven algorithms are proposed in MANET network. The demand-driven algorithms also can be seen as reactive routing protocols. According to Zhou (2003) claimed that the routing can divide into two parts in reactive routing protocols including route discovery and route maintenance. The route discovery is built up between source and destination node through broadcast a route discovery packet. It similar with an ARP request gets the MAC address of destination in wired network . The route maintenance means that check and repair the validity of the route on time due to the MANET network much more mobility than wired network. So the reactive algorithms are more scalable because of they reduce routing overhead . There are two typical reactive algorithms addressed in MANET network: DSR  and AODV .
The Dynamic Source Routing protocol (DSR) is efficient routing protocols which allows network self-configuring and self-organizing without any infrastructure . It is also consisted of two steps of route discovery and route maintenance.
Route discovery: If the destination address not find in routing cache, a route request packet is broadcast in the MANET network. And intermediate nodes' IP address will be stored in route request when it broadcast . When the destination nodes receive the request, they will accumulate the path. The result of route discovery is built up between source node and destination node and discovery the path to store in route cache.
Route maintenance: The DSR uses the DSR-specific acknowledgement to detect the errors in network communication . If the errors are detected the route discovery will re-initiate from source node.
So they are working together allow nodes to discover and maintain source routes to mobility target nodes in the MANET network . On the other hand, the DSR supports multi-paths. If one path failure, another path will implement in the communication and store into routing cache. So the DSR algorithms are more scalable because of they reduce routing overhead, even with continuous and quick move of all nodes in the network .
The Ad-hoc On-Demand Distance Vector Routing (AODV) is base on distance-vector routing algorithm. It is also suitable for a dynamic self-configuring and self-organizing network and similar to DSR, but DSR uses source routing. The AODV use the 'HELLO message' to determine local connectivity. Since it reduce response time to routing requests. Moreover, the sequence numbers are assigned to routes and routing table entries. So if any errors are detected by Route Error packet (REER), the nodes just maintain the sequence number and broadcast ID . Therefore the AODV can establish quick connection between nodes without any further delay . On the other hand, the AODV support the multicast transmission, it becomes important for multimedia communication in the future . The implementation of multicast is similar with unicast. The group leader node of AODV will broadcast 'GROUP HELLP message' throughout the MANET network and every node also are assigned sequence number. The procedure of multicast is that every time a node joins into a multicast group, it will send a request packet (RREQ) to multicast destination. Then the destination will send back a reply packet to source node, while this process is unicast, in order to prevent the loop all the time. Therefore the AODV supports high congestion and heavily traffic network of successful delivery .
In sum up, the DSR and AODV use lots of intelligence technical to achieve good performance and operation for the MANET network. For instance, they are both reduce the bandwidth and suitable for a dynamic self-configuring and self-organizing without any infrastructure  network. And the AODV support the 'HELLO message' and sequence number which can quick establish between each node without delay. Moreover, the multicast technical help network to delivery heavily traffic network. A [Table 1] shows the summary of comparison of DSR and AODV.
3.3 Hybrid algorithm
In practice, a number of routing algorithms are hybrid such as ZRP  and ZHLS and others. Due to this routing algorithm combine the advantages of proactive and reactive protocols. The document focuses on Zone Routing Protocol (ZRP) that consists of three routing protocols including Intrazone routing protocol (proactive protocol), Interzone routing protocol (reactive protocol) and Bordercast resolution protocol . Each routing protocol gives ZRP different functions, the Intrazone routing uses to maintain the local topology , the Intrezone routing finds optimal path between source nodes to destination nodes  and the Bordercast routing reduce transmission errors when find optimal path (router discovery) . Since the ZRP combines all intelligence technical from proactive and reactive routing protocols. The [Table 2] shows a performance of comparison of these three routing protocols.
From the table, we can see that the hybrid routing protocols give the MANET lots of benefits such as delay and overhead of route discovery in reduced and proactive routing protocols can help the route maintenance of reactive routing protocols and so on. Therefore the hybrid routing protocols are much better than another two routing protocols.
In this document we have presented some challenges of MANET network such as the network mobility and dynamic and lack of infrastructure. It leads to the topology changes rapidly and unpredictably. All these challenges promote the routing algorithms must be robust and adaptive in a self-configuring and decentralized network. Therefore the MANET networks adapt intelligence technical to achieve good performance for network communication including find the optimal path, minimise latency, minimise packet loss and reduce failure in transmission and others . The OLSR address Multipoint Relaying (MPR) flooding technique , Dijkstra's shortest-path algorithm and broadcast route discovery mechanism to achieve good performance for the MANET network. For instance, the MPR technical can help reduce flooding node information, reduce the number of topology broadcast packets and reduce the number of retransmissions packet. Therefore lots of bandwidth is saved in communication process. Moreover, the AODV support the 'HELLO message' and sequence number which can quick establish between each node without delay, and the multicast technical help network to delivery heavily traffic network. In addition, the ZRP combines all intelligence technical from proactive and reactive routing protocols. So the routing protocols incorporating intelligence can improve all these challenges and its operation.