Intelligent Water Drop Based Qos Computer Science Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Mobile ad-hoc network operates with no pre-setup infrastructure and is one of the most active research areas. In this kind of network mobile nodes are completely independent, self managed and they are highly dynamic in nature. Therefore, traditional routing cannot work properly in this environment. Besides this, Quality of Service (QoS) of the network is very important for real time and multimedia applications for providing better throughput. But providing QoS in routing is a challenging task. Thus in this paper we introduce a novel QoS aware multipath routing algorithm IWDRA, which is based on Intelligent Water Drop (IWD) algorithm and here packets follow the basic IWD properties among neighbor nodes. It provides better QoS of network which will increase network lifetime, network stability, packet delivery rate and it is also a highly adaptive routing which will support dynamic topology.


MANET [1] is a collection of mobile nodes which communicate with each other by multi-hop links and the transmission medium is radio wave. In this network any mobile node communicates directly with the other mobile nodes which are in the transmission range using the radio wave and uses intermediate nodes to communicate with other nodes which are out of its range. Routing in MANET is a challenging task because here topology changes dynamically due to nodal mobility. Thus efficient routing protocols are needed for transmitting data from one node to another node. There are mainly three types of routing protocols. In proactive routing, a continuous traffic monitoring is being done to know the topological information among the network nodes. In reactive routing routes are established only when needed and hybrid routing which is a combination of the above two types of routing. Now a days, the demand of real time and quality of service (QoS) requirements increases drastically. QoS in the network aims to find routes that can provide the required quality needed for the applications. QoS routing is a technique that takes into account the appropriate link information and based on that statistics it selects the path that satisfies the QoS requirements. QoS routing is key part of a QoS mechanism as its main function is to find nodes that satisfies the applications requirements. Providing QoS in routing in MANET is a critical task because of its dynamic nature. Various routing protocols have been designed to tackle with the problem of QoS aware routing. In this paper, a new QoS aware routing protocol has been proposed which is based on the dynamic of river systems, actions and reactions which occurs among the water drops in river. The natural water drops are used to develop IWD and a better solution of the problem is reached with the cooperation of each IWDs. This algorithm is mainly a population-based constructive optimization algorithm. In our routing protocol each packet has the IWD properties and it takes the QoS metrics energy, buffer space to increase the network stability and packet delivery rate respectively. Besides this, delay, bandwidth is also considered to reduce the end to end delay and utilize the link capacity properly. The main feature of IWD is its velocity. The link which has better quality, the IWD packet of travelling through that link will gain more velocity than the other IWDs and will reach its goal faster than other IWDs. This helps in quick convergence to the better solution. Thus this protocol can find better route in less amount of time. Also, here an efficient route failure management technique is used to increase network throughput and lifetime of network which is very important for real time and multimedia applications.

Exiting works on Swarm based routing in MANET:

Swarm intelligence is defined as the collective behavior of decentralized and self organized group. Here simple agents interact with the environment (called stigmergy) and between each other. The agents follow simple routes and possess themselves limited capabilities. They don't follow the centralized control for each individual and interact locally and randomly with each other. In global point of view, their collective behavior emerges towards an intelligent routing. The nature of swarms highly resembles the mobile ad hoc networks and thus the ideas from them are used to deign many intelligent routing algorithms. There are many swarm based routing algorithms exists. Some of them are described briefly. At first, Ant Colony Based QoS Routing Algorithm for Mobile Ad Hoc Networks [5] is an on-demand QoS routing algorithm proposed by P.Deepalakshmi et al. This algorithm is highly adaptive in nature and mainly reduces the end to end delay in high mobility cases. This is a good routing scheme when node mobility is high. Here in route discovery the minimum QoS requirements in terms of bandwidth, delay and hop count are considered. But the other QoS constraints i.e. other network layer or link layer metrics like energy, jitter, link stability etc. are not considered here. Next, an Ant Colony-based Multi Objective Quality of Service Routing for Mobile Ad hoc Networks [6] proposed by P. Deepalakshmi et al. is an ant-based multi-objective on-demand QoS routing algorithm for mobile ad hoc network. Here, Link failure is detected quickly as node uses updated view of the network with positive and negative feedback. As it is a multi-path routing algorithm, therefore, it supports node mobility in a better way. The main drawback of this approach is, at the time of route discovery, it only considers delay, throughput and jitter as the QoS metrics but other network layer or MAC layer metrics are not considered for achieving more stable route and high throughput. Next, S. Kanan et al. proposed Ant Colony Optimization for Routing in Mobile Ad-Hoc Networks [7] which is a multi agent ant based routing algorithm for MANET. This is a new technique increases the connectivity of the nodes during high node mobility and for this the QoS constraints like delay is decreased and packet delivery ratio (PDR) is increased. It mainly focuses on the node mobility by only supporting QoS requirements like delay, PDR etc. which is not sufficient for multimedia application. Other network and link layer metrics have to be considered. There are various other issues. This algorithm uses both proactive and reactive routing. Paths are monitored always, so the overhead is increased in routing. It does not take into account the heterogeneous MANET behavior, where the data rate of each link can be different, so flow control is necessary, which is not noticed here. Consequently, there is a lack of Congestion control. Moreover, in MAC layer as nodes communicate through the shared medium; hence, there is a chance of collision, contention and interference. These problems are not taken into account in this algorithm. Furthermore admission control is not tackled properly for better throughput and efficiency. S.B.Wankhade et al. proposed an on-demand routing algorithm, Route Failure Management Technique for Ant Based Routing in MANET [8]. This algorithm is inspired by the ant colony routing algorithm. Route failure management is the main key of this algorithm. Authors have shown that it has good maintenance scheme and it supports good packet delivery ratio (PDR) with less packet drop and delay. In this technique, at the time of route discovery, only delay, bandwidth, hop count, PDR are taken into account but other network layer metrics or link layer metrics like throughput, jitter, node buffer size etc. are not taken into consideration. Here, in route updation phase, multiple routes are found and in maintenance phase, fuzzy logic is used to find Link Stability Coefficient. So, a small overhead is involved with route updation in some time interval. Furthermore, here flow control is not handled explicitly, so when source generates HANT it does not takes into account the data rate of the link in heterogeneous MANET. Packet collision and channel interference also are not considered. The computation of fuzzy logic in the node may cause premature death to the nodes due to the limited energy of the nodes. For real time and multimedia communication only improvement of packet delivery ratio is not sufficient, because end to end delay, processing delay at each node and also other QoS constraints affects the throughput. A detailed literature survey on ant based QoS aware routing and their comparative analysis in MANET can be found in our previous work [9]. Next, Wedde et al. proposed a new routing scheme BeeAdHoc: an energy efficient routing algorithm for mobile ad hoc networks inspired by bee behavior [10]. This is an energy efficient routing algorithm and concept is taken from the foraging principles of honey bees. It is a reactive routing algorithm and mainly utilizes two types of agents- scouts and foragers for routing in MANET. This is a multipath routing algorithm which helps in effective load balancing and less congestion in the network. Moreover according to the authors this protocol consumes less energy compared to DSR, AODV and DSDV routing because it utilizes less control packets. Another bee inspired routing protocol is QoSBeeManet: A new QoS multipath routing protocol for mobile ad-hoc networks [11] proposed by S. Bitam et al. This algorithm is based on autonomic bee communication principle. Stochastic broadcast of scout packet is used to reduce the network congestion. Here, source node evaluates all discovered path for fairly distribution of data packets through the path. Authors show that QoSBeeManet has improved average bandwidth and end-to-end delay compared to DSR and AODV routing protocol. Next, Sharvani GS et al. proposed Adaptive Routing Algorithm for MANETs: TERMITE [12]. It is a adaptive, distributed, mobile agent based algorithm which was inspired by recent ant colony metaphor. In this algorithm a group of mobile agents called artificial termites build paths between pair of nodes. Here, the networks concurrently and information is exchanged to update the routing table. It also provides better load balancing as it is a multipath routing.

Intelligent Water Drops:

The IWD algorithm was first introduced by Shah-Hosseini in 2007.The IWD [2] possesses same functionality as natural water drops flows in the beds of rivers. This intelligent water drop, IWD for short has two important properties:

The amount of soil it carries, soil (IWD).

The velocity it possesses, velocity (IWD).

For each IWD, both property values (i.e. soil (IWD), velocity (IWD)) values may change as the IWD flows in the environment.

It is assumed that the environment where the IWD flows is discrete. The environment is represented by a graph (N, E), where N is the number of nodes and E is the number of edges in the graph. Every node in the network is linked by an edge which holds an amount of soil. Based on the conditions of the IWD flowing in the network, the soil of each edge may be increased or decreased.

Suppose, an IWD is in node S at time t and it wants to move to the next node K in the network. The amount of soil on the edge between these nodes denoted by soil (S, K) used for updating the velocity of IWD denoted by:

Here, is the nonlinearly proportionality. One possible formula, according to [2] is given in eq. (4.2) in which the velocity of the IWD denoted by is updated by the amount of soil, soil(S, K), between the two locations S and K:

vel IWD_Req (t+1) = vel IWD_Req (t) +

Here, av, bv, cv and α are constant velocity updating parameters that are set for a given problem. The updated velocity of the IWD, velIWD(t+1) after reaching node j will be equivalent to velIWD(t) + velIWD(t).

The amount of soil carried by IWD, soil (IWD) is increased over time by removing some soil from the path joining nodes S and K. The amount of amount of soil added to IWD, (IWD) and the soil removed from the path, (S, K) is inversely and nonlinearly proportional to time needed for IWD to travel from S to K, time (S, K; IWD).

(IWD) = (S, K)

According to [ ] time (S, K; velIWD) is the time taken for the IWD with velocity velIWD to move from node S to K, (S, K) is the soil added to the IWD.

Soil (S, K) =

Here as, bs, cs and are user-selected positive parameters.

The time duration for the IWD is calculated by the simple laws of physics for nonlinear motion. So, the time taken for the IWD to move from node S to K is inversely proportional to the velocity of IWD, velocity (IWD) as well as distance between two positions, d(S, K).

One example according to [2] is given in equation below with which the time taken for the IWD to move from S to K with can be calculated.

time (S, K; IWD) =

Here, HUD (.,.) is the local heuristic function defined for a given problem and it is used to measure undesirability of an IWD to move from one location to another location.

So, the amount of soil that has been removed from the path by the IWDs is totally dependent on the velocity of moving IWD. Velocity is inversely proportional to the time taken to travel. Thus, the amount of soil removed is inversely proportional to the time IWD needs to pass from one node to another node. So, fast IWD will pick up more soil than the slower one.

Soil removed from the visited path between node S and K. The updated soil (S, K), is proportional to the amount of soil on the path joining S to K. Soil (S, K) = Soil (IWD), specifically

Soil (S, K) Soil (IWD)

According to [2], the updated soil on the path between S to K can be denoted by the following formula:

Soil (S, K) = Soil (S, K) - Soil (S, K)

Where and are offer positive numbers between zero and one. In [ ] = 1-

The soil of the IWD denoted by soil (IWD) is represented by:

Soil IWD = Soil IWD + Soil (S, K).

QoS metrics:


Delay is the latency consumed by a packet in MANET to reach from source to destination. The end-to-end delay of a packet depends on the processing delay, queuing delay, packetization delay of each node and transmission delay, propagation delay of each link of the path. So, end-to-end delay is the sum of delay created by links and nodes along the path. In ad hoc networks, propagation delays are negligibly small and almost equal for each hop in the path.


Bandwidth or available bandwidth is the QoS constrain which refers to the available capacity of a link to send packets from source to destination in the network. From this constraint the available data rate can be determined. Bandwidth estimation is important because each host has imprecise knowledge about the status of links and as network changes dynamically, bandwidth of the link also changes very frequently. For multimedia and real time applications, at the time of data transmission, choosing the link which has the bandwidth greater than or equal to the required bandwidth is very essential for better throughput.


In case of real time and multimedia applications, at the time of route discovery, a path is selected and it is maintained until the path breaks for uninterrupted communication. Hence, usage of the path for a longer period of time reduces the energy of the node along the path. It may also possible if the node belongs to multiple routes. Due to continuous transmission and reception of message, the battery power is drained out. When the battery power is low, the node cannot transmit or receive the message; as a result route break occurs. So, the protocol has to initiate another route discovery phase for finding another suitable route. So, this metric is taken into account which selects the node in the route having higher energy, to enhance the lifetime of the route.

Buffer size:

In MANET buffers are used in each node and there is a limit on the size of those buffers. Packet drop ratio decreases with the increasing size of buffer, thus increase the throughput. The increasing number of packets arriving to the node also increases the energy consumption by increasing the number of transmission and reception costs and decreasing average lifetime of the node. So, in this reason this metric is taken into account for efficient buffer management.

Hop count:

Hop count is the number of hosts or routers required to transfer data from source to destination. It is important because if the number of hop counts increase then delay will increase. As MANET is dynamic and due to nodal mobility, network topology also changes dynamically. So, for this if the paths with minimum hop count is needed, otherwise packet loss will increase drastically causing low throughput. In this reason, hop count is taken into account as a QoS constraint.

Mathematical Model

For mathematical analysis MANET is represented by a connected undirected graph. Let G (V, E) represents the mobile ad hoc network. Here V denotes the set of network nodes and E denotes the set of bidirectional links. QoS metrics with respect to each link e ∈E is delay (e) and bandwidth (e). With respected to node n ∈V , it is delay (n), energy (n) and drain rate (n) which is the energy dissipation rate of node 'n'. Another QoS metric considered here is hop count. It is important because multiple hops are used for data transmission in MANET. So, it is necessary to find paths with minimum hops. The main motivation of this proposed algorithm is to find path from source to destination which will satisfy the QoS requirements such as delay, bandwidth, energy, drain rate and hop count.

Let, path (i, j) or R is entire path from node i to j where QoS constraints have to satisfied.

From an arbitrary node i to an arbitrary node j, delay, bandwidth, energy, drain rate and hop count is calculated as-

delay (path (i, j )) or D (R) = +

where, delay (path (i,j )) is the transmission and propagation delay of the path(i,j) and delay (n) is the processing and queuing delay of node 'n' on path(i, j).

bandwidth(path(i,j)) or B(R)= {bandwidth(e)}

where, bandwidth (e) is the available bandwidth of that link on path(i, j).

energy (path (i, j)) or E (R) = { energy (n)}

where, energy (n) is the residual energy of node 'n' on path(i, j).

buffer size (path (i, j)) or BUFF (R) = {buffer size (n)}

where, buffer (n) is the available buffer space of node 'n' on path(i, j).

hop count (path (i, j)) or HC (R) = Number of nodes in the path.

Data Structures:

The following data structures are used for our propose IWDRA routing.

HELLO Packet:

HELLO packets are broadcasted periodically to the neighbor nodes to know the status of remaining energy and buffer space of the node. Based on the size of HELLO_Ant packet and starting, receiving time, current node will calculate the available bandwidth of the outgoing links. The format of the packet is given below:

0 2 9 17 31

Pkt_Type = HELLO


Buffer Space

Starting time

Fig.1. HELLO packet

IWD_Req Packet:

It consists of request starting time, amount of soil the packet carrying, source id, destination id, stack of visited node addresses. This packet is used in the route finding phase.

0 2 17 31

Source IDPkt_Type = IWD_Req

Req_Starting Time


Destination ID

Stack of nodes visited



Fig.2. IWD_Req Packet format

IWD_Rep Packet:

It consists of hop count, delay between two nodes, node's remaining energy, nodes remaining buffer space, available bandwidth, amount of soil the packet carrying, source id, destination id, stack of node address to be visited. This packet also used in route finding phase.

0 2 6 11 16 21 26 31

Pkt_Type =IWD_Rep





Buffer Space



Source ID

Stack of nodes to be visited



Destination ID

Fig.3. IWD_Rep Packet format

IWD_Update Packet:

It consists of request starting time, amount of soil the packet carrying, source id, destination id. This packet is used in the route finding phase for updating the status of the current link.

0 2 17 31

Source IDPkt_Type = IWD_Update

Req_Starting Time


Destination ID

Fig.4. IWD_Update Packet format

Error Packet:

This packet is used in route failure handling phase and it consist of the node id where the failure occurs and the source node id where the error packet is to be sent.

0 2 31

Pkt_Type = Error

Source Id

Destination Id

Fig.5. Error Packet format

Data Packet:

This packet is also used in route failure handling phase and it consist of the node id where the failure occurs represented by Destination ID 1, the node id where the data packet will be sent represented by Destination ID 2 and the source node id where the error packet is to be sent and data field.

0 2 31

Pkt_Type = Data

Source ID

Destination ID 2

Destination ID 1


Fig.6. Data Packet format

Node Cache:

This table is stored in each node. It consists of the unexpired path, the path preference probability of that path with respect to a destination, the updated value of the packet soil and the current time when the soil is updated.


Path Preference Probability

Updated Packet Soil


Fig.7. Node Cache format

Node routing table:

This table is also stored in each node. It consists of the neighbor node id, its remaining energy, buffer space and the link soil between them.

Node id


Buffer space

Link Soil

Fig.8. Node routing table format

Proposed Algorithm:

This proposed routing algorithm is a QoS aware multipath adaptive routing algorithm. It has three phases namely Route finding phase, Route maintenance phase and Route failure management phase. In route finding phase three packets are used namely IWD_Req packet, IWD_Rep packet and IWD_Update packet. In route finding phase, at first IWD_Req packet is send by the source node towards the destination through all its neighbors. At the time of traversing the link that packet updates the link soil and the packet soil by following the IWD properties. When it reaches to the destination it is converted to IWD_Rep packet and it is send towards the source node by following the same path as in IWD_Req packet. At the time of returning IWD_Rep packet collects the minimum bandwidth, minimum remaining energy of a node and minimum buffer space and with that information the node calculates the path preference probability of that path and if it satisfies the minimum QoS constraints that path is stored in the node cache with amount of soil and current time. When the packet reaches the source node the same procedure is followed and the path with better path preference probability is used for routing data. When data reaches the destination an IWD_Update packet is generated and is send to the source by collecting the link status and if it is degrading then the source node will choose another path with net better path preference probability for routing. This is taken care of in the route maintenance phase. Next, when link failure occurs there can be three possibilities. If the node where the failure occurs has another valid path to the destination then that path is used for routing. If the effected node has no valid path to the destination then it generated the Data packet and sends it to the neighbors. If the neighbor node has a valid path to the destination then it generates an Error packet and sends it towards the source by deleting the invalid routes from the cache as well as sends the undelivered data to the destination node. When source node receives the Error packet if it has a valid path to the destination, that path is used for routing otherwise a new route finding phase started. Lastly, if a link failure occurs and there is no valid path to the destination then the node generates the Data packet and sends it towards the destination through all its neighbors and all invalid paths are deleted from the cache. When source node receives the Data packet it finds that there is no path to the destination, all timers are expires and network partition detected. At that time it starts new route finding phase.

Route finding phase:

Step 1: Suppose Source S wants to send data to destination D with QoS constraints delay, bandwidth, buffer space, energy and it has no valid path exits in the cache.

Step 2: Source node creates an IWD_Req packet which and sends the packet to its neighbor node by updating the soil and velocity of IWD_Req packet and soil of link between them. The soil and velocity is updated according to the given formula:

2.1. For each IWD_Req moving from node i to j the velocity vel IWD_Req is updated by:

vel IWD_Req (t+1) = vel IWD_Req (t) + (1)

Where vel IWD_Req (t+1) is the updated velocity of IWD_Req packet.

2.2. For each IWD_Req moving on path from node i to j, compute the Soil (i, j) that the IWD_Req packet loads from the path by:

Soil (i, j) = (2)

Such that

time)) = (3)


HUD (j) = (4)

Here λB, λE, λD and λBUFF are the weight factors which indicate the relative significance of the QoS parameters during soil and velocity update on path (i, j).

2.3. Update the Soil (i, j) of the path from node i to j traversed by that IWD_Req and also update the soil that the IWD_Req carries SoilIWD_Req by:

(5) Soil (i, j) = (1-). Soil (i, j) - . Soil (i, j)

SoilIWD_Req = SoilIWD_Req + Soil (i, j)

Step 3: When it reaches to the intermediate node it first check if it has a valid route in the route cache. If it has the valid route, then that path is selected for routing, otherwise, the IWD_Req packet is forwarded towards the neighbors according to the previous process.

Step 4: When the IWD_Req packet reaches the destination, it will convert to the IWD_Rep packet and uinicasted towards the source following the same path as in IWD_Req packet. In the returning phase IWD_Rep packet will collect the maximum buffer space and minimum energy of nodes and minimum bandwidth of the link in the path. Each node calculates the path preference probability by:

Pijd= (6)


Soil (i,j)=


Bijd= bandwidth (path(i,d))

Eijd = energy (path (i,d))

BUFFijd= buffer size (path(i,d)).

The node stores the path in the cache if it satisfies minimum Path Preference Probability.

Step 5: When IWD_Rep packet reaches the source node it will calculate the Path preference probability and the path which satisfies minimum QoS constraints will store in the cache and also the Soil (i, j) value of IWD_Rep packet is stored. The path with better Path Preference Probability is selected for routing data packets.

Step 6: Whenever data packet reaches the destination node, the node will create an IWD_Update packet and send it towards the source node for updating the link quality. The soil updation is done according to the formula (4).

Route Maintenance Phase:

When IWD_Update packet finds that the path preference probability of an intermediate node decreases, the value become negative in eq. (4) applied as previous, otherwise it remains same as before. When the IWD_Update reaches to the source the corresponding soil of the path and the time is updated. If the updated soil value remain negative or zero (e.g. updated soil= Previous soil-New soil 0) then no action taken and routing continued with existing path and otherwise a new unexpired path with next best Path Preference Probability is selected for routing.

Route Failure Management Phase:

Case 1: Intermediate node has a valid path to the destination

If an intermediate node detects a route failure, it will first search its cache for other valid routes, if it exists then the path with better path preference probability will be selected for routing. In parallel the node will send an error packet towards the source node which contains the information about invalid route and every node in the path along with the source node will deletes the invalid route from the cache and routing will continued with the new path. When source node receives the Error packet if it has a valid path to the destination, that path is used for routing otherwise a new route finding phase started.

Case 2: Intermediate node has no valid path to the destination

When source node sends data packet towards destination it will set a timer, and if an IWD_upadate packet does not come for that route within the time, then source will find that there is a problem with that route, i.e. there might be a link failure. At that time the source node will stop sending the data with that path and the path which is unexpired and has better path preference probability will be selected for routing. In parallel the intermediate node where the link breakage occurs will send a Data packet towards the source and whenever a node has a valid, unexpired path to the destination, it will extracts the data and will send it to the desired destination. It will create an Error packet from the information of the Data packet and will send that packet towards the source. Hence all invalid routes are deleted from the node cache in the path along with the source node.

Case 3: Network has no valid path to the destination

At this moment network partition occurs i.e. there is no valid path to the destination. The node where failure occurs will generate Data Packet and it will reach the source node and all invalid paths will be deleted and the timer of source node for each path will expire.


Case 1:

Suppose A is the source node and E is the destination node. During data transmission the link between node D and E fails which is shown in Fig 9. Then node D delete the invalid path and broadcasts an Error packet where the source id is D and the destination id is A. So, all invalid paths going through node D are deleted from the cache of nodes. It is shown in the Fig 10.

Fig.9. Link between D and E fails.

Fig.10. Every path going through node D is deleted from the cache of every node.

Case 2:

Suppose, link between node C and D fails. It is shown in Fig 11. There are no valid paths from C to the destination. So, C will generate a Data Packet and will send it to node B. As node B has a valid path to node E, it will extracts the data and send it to node E. As well as node E creates Error packet to the source node A and deletes invalid routes from the cache. This phenomenon is shown in Fig 12.

Fig.12. Node C sends Data Packet to B and sends the undelivered data to desire destination E as well as it sends Error Packet to A by deleting all invalid paths.

Fig.11. Link between C and D fails

Case 3:

In Fig 14, link between node E and F fails. There are no valid path exits to destination F. At this moment network partition occurs. So, E generates Data packet and broadcasted towards the source node A by deleting all routes from the cache.

Fig.13. Network configuration after node F joins the network.

Fig.14. Network partition occurs. Node E sends Data Packet towards A and all invalid routes are eleted from the cache.

Performance Analysis:

The main goal of this algorithm is to find a stable route from source to destination which will satisfy QoS constraints and will support multimedia, real time applications.

1. This is an adaptive routing algorithm and suitable for the network where mobility is high.

2. This algorithm finds multiple paths from source to destination which will satisfy QoS constraints delay, bandwidth, energy, buffer space and hop count.

3. Here, energy and buffer space of individual nodes are considered as QoS metrics which will increase the network lifetime and packet delivery rate respectively.

4. In this routing if a node does not satisfies the QoS constraints, then IWD_Rep packet is not forwarded further and thus it reduces the number of control packets that helps to utilize the bandwidth properly.

5. The best feature of IWD is its velocity. The convergence of optimal path with satisfying QoS constraints can be found in finite and less amount of time as the velocity of IWD of the path with better path preference probability will be high and it will reach faster to the source node by making quick convergence to the better route.

6. In this routing, IWD_Update packet is used to continuous monitoring the quality of the active route an whenever a change occur it can detect quickly. If the quality is degrading it can use another path with next better path preference probability for routing which will reduce the packet loss rate.

7. When route failure occurs the IWD_Update packet will not reach to the source node in time and the timer expires, so quickly the source node can select another better path for routing without waiting for the error packet. It increases network lifetime and packet delivery rate as well as throughput of the network.

8. It is a loop free routing and it uses route required flag. So end to end delay, packet loss rate decreases and throughput of the network increases.


In this paper a novel paradigm for QoS aware routing in MANETs based on the properties of intelligent water drop is proposed. IWDRA is an adaptive multipath routing algorithm and it is suitable in case of high degree of mobility. This algorithm is aware of node's remaining energy as well as buffer space which is very useful for increasing network lifetime and successful delivery of data packets. As it is a multipath routing bandwidth is utilized properly by enhancing efficient load balancing capability. It also reduces end to end delay at the time o routing. In this routing, at first for the velocity of the IWD packets the algorithm converges towards the proper path quickly which will satisfy the required QoS constraints and it reduces the route finding time. Secondly, in route failure management phase, a mechanism is used through which network converges towards the stable sate quickly and in a finite amount of time, which reduces the packet loss, increases network stability which is very important for real time and multimedia applications. Besides this, in route maintenance phase during data transmission always a route is monitored such that any degradation of quality of the route cannot effect the routing which also increases overall throughput of network.