a high performance scalable qos routing protocol for ... high performance scalable qos routing...

A High Performance Scalable QoS Routing Protocol for Mobile Ad Hoc Networks

Bita Safarzadeh

Department of Computer Engineering

Shahram Jamali Hamed Alimohammadi

Islamic Azad University

Ardabil Branch

Ardabil, Iran [email protected]

Faculty of Technical and

Engineering University of Mohaghegh

Ardabil, Iran

[email protected]

Department of Computer Engineering

Ardabil BranchIslamic Azad University

Ardabil, Iran [email protected]

Abstract A mobile ad hoc network (MANET) consists of a set

of mobile hosts capable that form a temporary network on the fly without using any fixed infrastructure. The nodes operate both as hosts and routers. Due to mobility of nodes, the topology of the network may change rapidly and unexpectedly, thus setting up routes that meet high reliability is very challenging issue. Hence it is very important to find a route that can obtain highly stable routing and transmission ratio during data transmission.

To address the above mentioned problem, we propose a high performance scalable QoS routing protocol, which bases on the route life time that is obtained using Global Positioning System (GPS), the residue energy and hop count. Therefore in this scheme, data is sent through a route with high stability, high residue energy level and low latency. The protocol consists of three phases: route discovery, route selection and route maintenance. Simulation results the HPSQR protocol achieves high performance and scalability with high packet delivery ratio and throughput as compared to best-known on-demand protocol, AODV.

1. Introduction

A mobile ad hoc network (MANET) sometimes called a mobile mesh network, is a self configuring network of mobile devices connected by wireless links. Each device in a MANET is free to move independently in any direction, and will therefore change its links to other device frequently. Each most forward traffic unrelated to its own use and therefore be in router. The primary challenge in building a MANET is equipping device to continuously maintain

the information required to property route traffic. Many routing protocols have been proposed for MANETs [8,14,15]. These routing protocols can be classified to two classes: Table-driven (also called Proactive) and On-demand (also called Reactive). The proactive routing protocols maintain fresh lists of destinations and the routs by periodically distributing routing tables throughout the network. Main disadvantages of such algorithms are: Respective amount of data for maintenance, slow reaction on restructuring and failures. Well known protocol of this type is the DSDV [14]. The main disadvantages of such algorithms are: high latency time in route finding, excessive flooding can lead to network clogging. The typical on-demand methods are the ad hoc on-demand distance vector routing (AODV) [15] and the dynamic source routing (DSR) [8].

While DSDV itself does not appear to be used today, other protocols have used similar technique. The best-known sequence distance vector protocol is AODV, which, by virtue of being a reactive protocol, can use simpler sequencing heuristics. In DSR is similar to AODV in that it forms a route on-demand when a transmitting node request one. However, it uses source instead of relying on the routing table at each intermediate node. The mobile ad hoc wireless networks, are rapidly becoming the preferred solution for flexible and low cost networking. In addition, due to its flexibility, this scheme is often deployed in critical applications that need exact data collection such as military and first responder applications. Most MANET routing protocols focus on finding a feasible route path from a source node to a destination node without any consideration for utilization of network resources or for supporting specific application requirements, such as Quality of Service (QoS) services [8,21].

Bita Safarzadeh,Shahram Jamali,Hamed Alimohammadi, Int. J. Comp. Tech. Appl., Vol 2 (2), 332-339

332

ISSN:2229-6093

RAM

Typewritten Text

RAM

Typewritten Text

RAM

Typewritten Text

RAM

Typewritten Text

RAM

Typewritten Text

RAM

Typewritten Text

RAM

Typewritten Text

Many applications of MANETs need reliable packet delivery. Therefore, in such applications most important parameter of Qos is the Packet Delivery Ratio (PDR). For maximizing this parameter, routing of network must find routes with high stability and sufficient energy level.

In this paper, we propose high performance scalable QoS routing protocol (HPSQR) for MANETs by constructing most reliable route path from a source node to a destination node. Our protocol is an enhancement over AODV. It selects a routing a path based on two following parameter:

1. Stability of paths that found from the source node to the destination node. For calculating stability of a path, HPSQR uses the route life time (RLT) between two connected mobile nodes by using global positioning system (GPS).

2. Residue energy of paths that from the source node to the destination node. For determining residue energy of a path, we find node with minimum residue energy along that path.

Then, we use three parameters, the route life time, the residue energy and the number of hop, to select a routing path with high stability, sufficient energy level and low latency consequently high PDR.

2. Related work

A mobile ad hoc network is a dynamically

reconfigurable wireless network that does not have a fixed infrastructure. Due to the high mobility of nodes, the network topology of MANETs changes very quickly. Therefore, finding the routes that message packets use becomes more complex. QoS routing supports reliable applications such as military and responder applications. The protocol must establish a route that satisfies certain QoS constraints, such as delay, power and throughput.

Using different techniques result in different constraints on data transmission. Some researchers have proposed proactive routing protocol for QoS support [4,18]. However, the proactive protocols are more reliable to suffer performance degradation than on-demand protocols due to state route information. Many on-demand QoS routing protocols have been proposed for MANETs. In [1], the genetic algorithm was used to perform faster route discovery. In [6], bounded flooding was proposed to increase the chance of finding feasible branches. In [19], the adaptive QoS protocol was proposed based on the prediction of local performance. In [24], QoS routing in IEEE 802.11b was proposed to calculate bandwidth and end-to-end delay for routing. In [12], a subset of end-to-end paths was used to increase stability and reliability of the route. In [3], was extended the AODV routing protocol

for MANETs to a multi-path protocol with consideration of end-to-end delay for route selection.

[17], presented multi-path routing protocol that considered the two issues of reliability and security for QoS support and quantified the security of the protocol in terms of the number of eavesdropping nodes. In [25], was proposed a solution based on the swarm intelligence paradigm for QoS support.

Recently, many routing protocols were proposed for MANETs that use global positioning system [2,23]. The coordinates of each node can be known using GPS. Further, the transmission routing protocols can complete the process of route discovery by mathematically calculating the routing. Due to mobility of mobile nodes in MANETs, the shortest path is not necessary the best path. If we do not consider the stability of routing path, the wireless links may be easily broken. There have been many efforts made to design a reliable routing protocol to enhance a network’s stability [5,7,10]. When a route is broken during routing, route discovery and maintenance procedure are executed. However, these procedures use many resources. To minimize route breakage, it is important to find a route that endures longer [11,20]. If a route discovery algorithm can find a stable route that endures longer, reduce route discovery packets and route maintenance overhead can be reduced.

[22], proposed a reliable multi-path QoS routing protocol with a slot assignment scheme. In this scheme, the QoS routing protocol is associated with searching for a reliable multi-path (or uni-path) Qos rote from a source node to a destination node in a MANET. In the protocol two parameters, the route life time between two connected mobile node and the number of hops, use to select a routing path with low latency and high stability. In [13] was proposed a new multi-path distance vector routing protocol. This reactive algorithm computes alternate routes loop-free and zone-disjoint paths.

Many algorithms have been were proposed, but we can say that the AODV protocol is must employed routing protocol in MANETs. In order to make reliable and scalable the AODV algorithm, we should take into account more parameters in route selection, in addition to hop count, because the shortest route is not always the best route. For this goal, the residue energy and stability are two most important concepts.

3. High performance scalable QoS routing protocol

In a MANET, the source node that needs network information for routing refreshes its


333

ISSN:2229-6093

information frequently due to ability of nodes to move at any time and changing network topology. Furthermore, the MANETs are constrained in many fields such as power and link life time. Thus, setting up routes that meet high reliability is very challenging issue. Hence, it is very important to find a route that can not only satisfy the QoS guarantee, but also obtains highly stable routing and transmission ratio during data transmission. In this paper, we propose a high performance scalable routing protocol with mobility prediction for MANETs. The proposed protocol includes route discovery, route selection and route maintenance. The route maintenance in this protocol works same to the route maintenance in the AODV protocol.

3.1 Link life time

In this section, we introduce the mobility prediction

method. This method uses the location obtained by GPS. For a mobile device, such as a PDA or a notebook, having a GPS receiver, longitude and latitude can be obtained. By matching the known longitude and latitude to the map, we can obtain the position. By continuously updating position, a GPS receiver can also provide data regarding speed and direction of travel. We assume that two nodes A and B are within the same transmission r of each other. We let (푥 , 푦 ) be the coordinate of mobile node A and (푥 , 푦 ) be the coordinate of the mobile node B. we let 푣 and 푣 be the mobility speed and 휃 and 휃 (0 ≤ 휃 , 휃 ≤ 2휋) be the direction of motion. Thus, we can obtain the link life time (LLT) using [16]:

퐿퐿푇 = −(푎푏 + 푐푑) + (푎 + 푐 )푟 − (푎푑 − 푐푏)

(푎 + 푐 ) (1) Where

푎 = 푣 cos 휃 − 푣 cos 휃 , 푏 = 푥 − 푥 , 푐 = 푣 sin 휃 − 푣 sin 휃 , 푑 = 푦 − 푦 .

When a source node sends a route request packet,

the node location, direction and speed is appended to the request packet. The next hop of the source node receives the request packet to predict the link life time between itself and the source node. If node B is the next hop of the packet for node A, node A inserts its location information in the request packet. Thus, node B can calculate the life time between itself and node A.

3.2 Route discovery

The discovery process is initiated whenever a source node attempts to communicate with another

node for which it has no routing information. Like ad hoc on-demand distance vector routing, our protocol neither maintains any routing table nor exchange routing table information periodically. When a source node requires a communication, it broadcast a route request (RREQ) packet to its neighboring nodes until they reach to the destination node. In our protocol, each RREQ packet records information of all links-status and nodes along traversed path. Link-status information is delivered from the source node to the destination node. The destination node may receive link-status information from different RREQ packets, it is mean that each RREQ packets traverses different paths. Finally, required computation for route selection is done at the destination node and result is backward to the source node, in order to decide route. We define RREQ packet format as follows:

(1) Type: the type of packet; (2) S: the source node address; (3) D: the destination node address; (4) SEQ: the packet sequence number; (5) LI: the location information of mobile nodes that includes location, velocity and direction; (6) RLT: minimum of link life time of links that RREQ packet traversed them; (7) RME: minimum of residue energy level of nodes that RREQ packet traversed them; (8) HC: hop count; (9) TTL: the limitation of hop length of the search path.

Now, we explain route discovery process. To

start we should describe four parameters: LLT: denotes the link life time between two

nodes, which is calculated by using equation (1). Each mobile node indentifies its location information using GPS. First, when a source node broadcasts a RREQ packet, it appends its location information (such as location, velocity and direction) to the RREQ packet. Upon a node receives the RREQ packet, it calculates the link life time by using location information in the RREQ packet and its own location information.

RLT: is the minimum link life time along a routing path. Therefore, the RLT is equal to the minimum of LLTs for a route. When a node receives the RREQ packet, it obtains the minimum of the LLT in the RREQ packet and the LLT calculated in this node then appends this value to the RLT field of the RREQ packet. Thus, the destination node receives the RREQ packet that contains the RLT.


334

ISSN:2229-6093

RME: is the minimum of the residue energy level of nodes along a routing path. When the source node broadcasts the RREQ packet, the nodes that in its transmission range, receive these broadcasted packet. Then, these nodes append their own residue energy level to the RME field of the RREQ packet. When a node receives the RREQ packet, it checks that if the destination node address is equal to its own address. If not, it obtains the minimum of the RME field of the RREQ packet and its own residue energy level, and then appends its value to the RME field.

Hop count (HC): is number of links of a path. Figure (1) shows the RREQ packet delivery

procedure. When a node wants to send some data to a destination node, it searches its own routing table. If there is a route to the destination node, it sends data through that route otherwise starts route discovery process and broadcasts a RREQ packet. When a node receives the RREQ packet, it searches its routing table. If there is a route to the destination, it replies to the source node. Otherwise, it broadcasts received RREQ packet to its neighboring nodes.

Figure 1. The RREQ packet delivery

Now, we describe stages of the RREP packet

delivery. Stage 1: Source node S sets up TTL to number of

all the nodes in the networks. Then it broadcasts a route request packet to its neighboring nodes.

Stage 2: When the intermediate node N receives the RREQ, if it is not the destination, it updates some fields of the RREQ packet, such as LI, RLT and RME, as above explanations. Then it decrements the TTL value and increments hop count. If the node N is not the destination and also the TTL value is not zero, then it forwards the RREQ packet to its all the neighboring nodes. If TTL value is zero, it drops the RREQ packet and does not re-forward that to any node.

Stage 3: If node N receives a RREQ packet and it is the destination node, it first appends the RREQ packet information to its own routing table and after that, calculates R value to find the ratio of traversed path. Then, it appends this value to the route reply (RREP)

packet and backwards the RREP packet to the source node through same path. The destination node does this operation for all of received RREQ packets. The R value is main parameter for route selection in the HPSQR routing protocol. The R is defined as follows:

푆 = 푅퐿푇 × 푅푀퐸

퐻퐶 (2)

Clearly, in order to maximize the R value, RLT

and RME should be maximized and HC should be minimized as possible. Actually, smaller hop count leads to route stability, it means higher reliability. And smaller value of R indicates higher mobility of each node in the path, it means lower reliability. We define the format of the RREP packet as follows:

(1) Type: the type of packet; (2) S: the source node address; (3) D: the destination node address; (4) SEQ: the packet sequence number; (5) R:the ratio of path; (6) TTL: the limitation of hop-length.

Stage 4: After sending the RREQ packet, the

source node waits a period of time. Consequently, it may receive different RREQ packets from different routes. At the end of the period of time, it finds the maximum of the R values of all received RREQ packets and selects the path with maximum R value for sending data, and saves this value in its routing table. Therefore, data is sent through a route with low latency, high stability and sufficient residue energy level. In fact, there is a trade-off between these three concepts in proposed protocol.

In the following, we give an example of the routing process for proposed protocol. In Figure (1) a part of a network is shown as example. If the source node S wants to send some data to the destination node D, it broadcasts a RREQ packet with the destination address equal to the node D address. Firstly, neighboring nodes of S receive the RREQ packet (A and B) and append required information to it, then forward to their neighboring nodes. This process will be continued until the RREQ packet arrives to the destination node.

As we described, the RREQ packet collects required information at each node at its way. Node D receives six RREQ packets from six paths. According to figure (1), these six paths are: path1 (S→A→B→D), path2 (S→A→C→D), path3 (S→E→F→G→D), path4 (S→E→F→C→D), path5 (S→E→F→C→A→B→D) and path6


335

ISSN:2229-6093

(S→A→C→F→G→D). In figure (1), the residue energy of each node is represented above the node and the LLT of each link is represented above the link. Based on this information, the node D calculates the R values. The R values of mentioned paths are: R1=49.8, R2=55, R3=63.75, R4=43.75, R5=20.85 and R6=33. Then the destination node D appends these values to the RREP packets and replies with the RREP packets along according paths.

Figure 2. Route reply process

Figure 3. Data packets routing process

Figure (2) shows the RREP packet sending

procedure. The source node S receives six RREPs after determined period of time for waiting for the RREP packets. Then, it selects path3 for sending data, because of its higher R value compared with other five paths. Thus, the most stable path in our protocol, HPSQR, is S→E→F→G→D. Figure (3) illustrates sending data packets from the source node S to the destination node D. Notice that in our protocol the shortest path is not necessary the best path because our routing protocol considers the stability and the energy concepts, as we saw in the example. 4. Simulation experiments

The benchmark protocol used to compare the

performance of HPSQR is the AODV [17]. The metrics used to assess the performance of HPSQR against the benchmark protocol are the packet delivery ratio and throughput. In section 4.1, we discuss the simulation environment and in section 4.2, we discuss the simulation results.

4.1 Simulation environment Extensive simulation experiments were

conducted using the Network Simulator version 2 (NS-2). NS2 is a scalable event-oriented simulation environment for wireless and wired communication systems.

HPSQR runs at the network layer using the MAC protocol 802.11. We assume that the mobility of the mobile nodes is random. In our simulation, each mobile node selects a position at random and moves to an arbitrary location at a randomly chosen speed. Node Mobility is changed by varying the node pause time. Pause time is inversely proportional to the mobility of the nodes, i.e., the lower pause time is the higher mobility of the node. The parameters used in our simulations are:

(a) Terrain dimensions: 1000 m × 1000 m (b) Simulation time: 1000 s (c) Mobility model: random way point (d) Pause time: 0-20 s (e) Speed of the mobile node: 0-5 m/s (f) Underlying MAC protocol: IEEE 802.11

We use CBR for the traffic model. Traffic model

parameters used in our simulations are:

(a) Packet size: 50 (b) Data sending rate: 1 Mbps (c) Time interval: 2s

4.2 simulation result

As mentioned earlier, the performance of

HPSQR was compared with AODV. AODV is a best-known on-demand routing protocol for ad hoc networks. We compare the performance of the protocol on the basis of the packet delivery ratio and throughput. We vary the number of nodes and the traffic load to change the degree of overhead of the routing and sending data in the network.

As mentioned above, we have different traffic loads in our experiments as follows:

(a) Low: 330 sent packets (b) Middle: 495 sent packets (c) High: 990 sent packets

In first step, we compare the HPSQR and the AODV in middle traffic load. Figure (4) shows the packet delivery ratio of the protocols as the number of nodes changes. We observe that the packet delivery ratio of HPSQR is considerably better than AODV. In figure (5), we see that the throughput of


336

ISSN:2229-6093

HPSQR algorithm is higher than of AODV. We see that when the number of nodes increases, throughput and also the packet delivery ratio of AODV decreases considerably, as it becomes increasingly difficult to find a better path for the source node, only with taking into account hop count. Against, the HPSQR protocol selects most reliable path at any situation of the network, due to considering the path stability and energy in addition to hop count.

Figure 4. Packet delivery ratio vs. number of nodes

Figure 5. Throughput vs. number of nodes

Figure 6. Packet delivery ratio vs. traffic rate with 200

mobile nodes Figure (6) and figure (7) show the packet delivery

ratio and the throughput of the protocols as the traffic load changes, respectively. These results obtained for constant number of nodes, 200 nodes. We observe that in the variant traffic loads, the HPSQR protocol works better than AODV, considerably.

Figure 7. Throughput vs. traffic rate with 200 mobile

nodes

Figure (8) and figure (9) present the packet delivery ratio and the throughput of the HPSQR as the number of nodes changes in the three different traffic loads. Results presented in figure (8) shows that the packet delivery ratio does not change meaningly. Therefore, we can say when the number of nodes and the traffic load increase at the same time, the packet delivery ratio increases also. Considering figure (9), we see that changes of the traffic load influence the throughput, higher traffic load leads to higher throughput. In addition, we can see the HPSQR protocol has excellent scalability at view point of the throughput; higher number of nodes does not decrease the performance.

Figure 8. Packet delivery ratio vs. number of nodes

with different traffic rate

Figure 9. Throughput vs. number of nodes with

different traffic rate

70

75

80

85

90

95

100

70 120 160 200 250

Number of Nodes

PDR

(%)

HPSQRAODV

0.17

0.175

0.18

0.185

0.19

0.195

0.2

0.205

70 120 160 200 250

Number of Nodes

Thro

ughp

ut (K

bps)

HPSQRAODV

70

75

80

85

90

95

100

Low Middle High

Traffic Rate

PDR

(%) HPSQR

AODV

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

Low Middle High

Traffic Rate

Thro

ughp

ut (K

bps)

HPSQR

AODV

70

75

80

85

90

95

100

70 120 160 200 250

Number of Nodes

PDR

(%) High

MiddleLow

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

70 120 160 200 250

Number of Nodes

Thro

ughp

ut (K

bps)

HighMiddleLow


337

ISSN:2229-6093

All of discussed results show that the HPSQR protocol is high reliable and scalable in different network situations, at least importing to two considered metrics. Analyzing the performance of HPSQR thought simulation we see that, it is an efficient algorithm whose packet delivery ratio and throughput are better than that of the AODV algorithm, due to selection best path at each time.

5. Conclusion

In this paper we proposed a high performance

scalable QoS routing protocol for MANETs. Using this approach we examined the QoS routing problem with searching for a high reliable route from a source node to a destination node. In order to select a reliable route, proposed protocol uses tree parameters, the route life time, the route minimum energy and the number of hops. Therefore, there is a trade-off between all of the important parameters in route selection. Through extensive simulation experiments, we observe that the HPSQR algorithm achieves high performance and scalability with high packet delivery ratio and throughput compared to the Ad hoc On-demand Distance Vector (AODV) routing algorithm.

References

[1] Barolli L, Koyama A and Shiratori N, “A QoS Routing

Method for Ad-Hoc Networks Based on Genetic Algorithm”, In: Proceedings of the 14th international workshop on database and expert systems applications, September 2003, pp. 175–179.

[2] Boukerche A, Rogers S, “GPS Query Optimization in Mobile”, In: Proceedings of the sixth IEEE conference on symposium, July 2001, pp. 198–203.

[3] Boshoff JN, Helberg ASJ, “Improving QoS for Real-time Multimedia Traffic in Ad-Hoc Networks with Delay Aware Multi-Path Routing”, In: Proceedings of the IEEE 2008 conference on wireless telecommunications symposium, April 2008, pp. 1–8.

[4] Chen S, Nahrstedt K, “Distributed Quality-of-Service Routing in Ad Hoc Networks”, IEEE Journal on Selected Area in Communications, 1999, pp. 1488–1505.

[5] Chiu C-Y, Wu H-K and Chen G-H, “Stability Aware Cluster Routing Protocol for Mobile Ad Hoc Network”, In: Proceedings of the ninth international conference on parallel and distributed systems. Chungli, Taiwan, 2002, pp. 471–479.

[6] Du X, “QoS Routing Based on Multi-Class Nodes for Mobile Ad Hoc Networks”, Ad Hoc Networks, 2004, pp. 241–254.

[7] Griffin TG, Bruce FB and Wilfong G, “The Stable Paths Problem and Interdomain Routing”, IEEE/ACM Transactions on Networking, 2002, p. 232–43.

[8] Hanzo L, Tafazolli R, “A survey of QoS Routing Solutions for Mobile Ad Hoc Networks”, Journal of IEEE Communications Surveys, 2007, pp. 50–70.

[9] Johnson DB, Maltz DA, “Dynamic Source Routing in Ad Hoc Wireless Networks”, Dordrecht: Kluwer Academic Publishers, 1996.

[10] Kim WI, Kwon DH ans Suh YJ, “A Reliable Route Selection Algorithm Using Global Positioning Systems in Mobile Ad-Hoc Networks”, In: Proceedings of the 2001 IEEE international conference on communications, 2001, pp. 3191–3195.

[11] Lim G, Shin K, Lee S, Yoon H and Ma J-S, “Link Stability and Route Lifetime in Ad-Hoc Wireless Networks”, In: Proceedings of the international IEEE conference on parallel processing workshops, 2002, pp. 116–123.

[12] Liu I-S, Takawira F and Xu H-J, “A Hybrid Token-CDMA MAC Protocol for Wireless Ad Hoc Networks”, IEEE Transactions on Mobile Computing, 2008, pp. 557–569.

[13] Loscri V, Marano S, “A New Geographic Multipath Protocol for Ad Hoc Networks to Reduce the Route Coupling Phenomenon”, In: Proceedings of the IEEE 63rd conference on vehicular technology, vol. 3, 2006, pp. 1102–1106.

[14] Perkins C, Bhagwat P, “Highly Dynamic Destination-Sequenced Distance-Vector (DSDV) Routing for Mobile Computers”, In: Proceedings of the 1994 ACM SIGCOMM Symposium on Communications, Architectures communications, architectures and Protocols, 1994, pp. 233–244.

[15] Perkins C-E, Royer E-M, “Ad Hoc On-Demand Distance Vector (AODV) Routing”, In: Proceedings of the 1999 IEEE workshop on mobile computing systems and applications, 1999, pp. 90–100.

[16] Rappaport TS, “Wireless Communications: Principles and Practice”, Upper Saddle River, NJ: Prentice-Hall, 1995.

[17] Reddy TB, Sriram S, Manoj BS and Murthy CSR, “MuSeQoR: Multi-Path Failure-Tolerant Security-Aware QoS Routing in Ad Hoc Wireless Networks”, Computer Networks, 2006, pp. 1349–1383.

[18] Sivakumar R, Sinha P and Bharghavan V, “CEDAR: a Core-Extraction Distributed Ad-Hoc Routing Algorithm”, IEEE Journal on Selected Areas in Communications, 1999.

[19] Sun H, Hughes HD, “Adaptive QoS Routing Based on Prediction of Local Performance in Ad Hoc Networks”, In: Proceedings of the 2003 IEEE conference on wireless communications and networking, 2003, pp. 1191–1195.

[20] Tseng Y-C, Li Y-F and Chang Y-C, “On route Lifetime in Multi Hop Mobile Ad Hoc Networks”, IEEE Transactions on Mobile Computing, 2003, pp. 366–76.

[21] Vergados DD, Vergados DJ, Douligeris C and Tombros SL, “QoS-Aware TDMA for End-to-End Traffic Scheduling in Ad Hoc Networks”, IEEE Wireless Communications, 2008, pp. 68–74.


338

ISSN:2229-6093

[22] Wang N-C, Lee C-Y, “A Reliable QoS Aware Routing Protocol with Slot Assignment for Mobile Ad-Hoc Networks”, Journal of Network and Computer Applications, 2009, p. 1153–1166.

[23] William S, Mario G, “IPv6 Flow Handoff in Ad Hoc Wireless Networks Using Mobility Prediction”, In: Proceedings of the 1999 GLOBECOM Conference, 1999, pp. 271–275.

[24] Xue Q, Ganz A, “Ad Hoc QoS On-Demand Routing (AQOR) in Mobile Ad Hoc Networks”, Journal of Parallel and Distributed Computing, 2003, pp. 154–165.

[25] Ziane S, Mellouk A, “A QoS Adaptive Multi-Path Reinforcement Learning Routing Algorithm for MANET”, In: Proceedings of the IEEE/ACS international conference on computer systems and applications, 2007, pp. 659–664.


339

ISSN:2229-6093

a high performance scalable qos routing protocol for ... high performance scalable qos routing...

Documents