ICT 2004

ICT 2004

PRDS: A Priority Based Route Discovery Strategy for Mobile Ad Hoc Networks Bosheng Zhou1, Alan Marshall1, Jieyi Wu2, Tsung-Han Lee1, Jiakang Liu1 1 Advanced Telecommunication Systems Laboratory, School of Electrical and Electronic Engineering, Stranmillis Road, The Queen's University of Belfast, BT9 5AH Belfast, Northern Ireland, UK 2 Research Center of CIMS, Southeast University, Nanjing China, 210096 Abstract-Most reactive route discovery protocols in MANETs employ a random delay between rebroadcast route requests (RREQ) in order to avoid “broadcast storms”. However this can lead to problems such as “next hop racing”, and high redundancy in the rebroadcasting. In this paper we propose a Priority-based Route Discovery Strategy (PRDS) to address both problems. PRDS is a distributed algorithm that assigns a high rebroadcast priority to “good” candidates for the next hop to reduce next hop racing; it also introduces a competing procedure to keep “bad” candidates from rebroadcast to alleviate “rebroadcast redundancy”. PRDS is a general mechanism that can be applied to different reactive routing protocols to improve the routing performance by enhancing route quality while reducing routing overhead. As an example, a macrobian route discovery strategy using PRDS (PRDS-MR) is introduced and evaluated via various simulations. Simulation results show that PRDS-MR outperforms AODV in terms of packet delivery ratio and delay while decreasing routing overhead up to 50%~80%. Index Terms-Ad hoc networks, PRDS, packet radio, routing, next-hop racing, rebroadcast redundancy, priority, wireless networks.

I. Introduction A mobile ad hoc network (MANET) is an autonomous system comprising a set of mobile nodes that can move around freely. Because MANETs don’t need any fixed infrastructure and can be fast deployed they have been attracting high interest in both military and civil applications. A MANET is generally formed as a multi-hop wireless network and routing plays an important role in its operation. Each node in the network acts as both a router and a terminal. MANETs are considered as resource limited, e.g., low wireless bandwidth, limited battery capacity and the like. Because of this, the conventional routing protocols used in fixed networks are no longer appropriate for MANETs due to the heavy routing overheads that consume too much bandwidth and energy. Existing MANET routing protocols may be classified into three categories: proactive, reactive, and hybrid. Proactive routing protocols are table-driven and rely on periodical exchange of route information. Each node maintains route entries to all other nodes of the entire network. In large and highly dynamic MANETs, Frequent routing information exchanges are needed to keep routing table up-to-date, and this leads to heavy routing overhead and large memory consumption. Reactive and hybrid routing protocols have been proposed to address these problems. In reactive routing protocols, each node only maintains active route entries and discovers routes only when needed. Routing overhead and routing table

storage can thus be reduced. In hybrid protocols, a network is partitioned into clusters or zones. Proactive and reactive routing protocols are then deployed in intra-cluster/ intra-zone and inter-cluster/inter-zone, respectively. The major advantage of hybrid routing is improved scalability; however, hierarchical address assignment and zoning/clustering management are complicated and can lead to heavy control overhead in high dynamic networks. The operation of a reactive routing protocol has three basic stages: route discovery, packet delivery, and route maintenance. Different reactive routing protocols are distinguished by different strategies taken in route discovery and route maintenance. Generally, route discovery is more costly in a dynamic network since it may need several route discoveries in a communication session because of network dynamics. In this paper, we focus on route discovery strategies for reactive routing protocols. In reactive protocols, route discovery usually works as follows: S1. Source S initiates a route request (RREQ) and broadcasts it to its neighbours. S2. On receiving a RREQ, each node rebroadcasts it. Each node only rebroadcasts the first copy of a RREQ so as to limit routing overhead. S3. The destination D sends a route reply (RREP) to S when it receives RREQ(s) destining to it. In step S2, each node usually rebroadcasts a RREQ in a random delay, e.g. in AODV [1] and DSR [2], so as to avoid a “broadcast storm” due to synchronization as identified by Ni et al [3]. We abbreviate this random rebroadcast delay route discovery approach as RD-random in this paper. Li et al [4] argued that RD-random might not find the most desirable route, and Zhou et al [5] demonstrated that flooding, which is a broadcasting scheme using random rebroadcast delay, cannot guarantee the least delay. Figure 1 illustrates a route discovery scenario. Two of the possible paths from Source S to Destination D are shown in the figure, i.e. path S-C-E-F-D and the shortest path S-A-B-D. Two problems exist if RD-random is used in this scenario: 1) Path S-C-E-F-D may be selected instead of the shortest path S-A-B-D by the destination D because the next hop of a constructing path in RD-random is randomly selected. This phenomenon was identified as “next-hop racing” problem in [4]. 2) All nodes except for the destination D will rebroadcast the RREQ. This is not a serious problem in this scenario; however it will lead to heavy routing overhead and consequent implications such as extra bandwidth and energy consumption in a large

C

S

E

A

G

F

D

B

H

I

Figure 1. A route discovery scenario. Source S tries to find a route to destination D. scale dynamic network. We identify this phenomenon as “rebroadcast redundancy” problem. A number of solutions have been proposed to solve the above problems individually. PANDA [4] uses different rebroadcast delay for good and bad candidates for the next hop, which are differentiated by positional attributes. ABR [6] and SSA [7] try to find stable or longer-lived routes by selecting stable links. DLAR [8], LBAR [9], and LSR [10] construct load balance routes by using some load metrics as route selection criteria. Works in [11,12,13] focus on energy-aware routing. Many techniques such as caching [2], query localization [14, 15], hybrid routing [16-18], exploiting location information [19,20], and Connected Dominating Set approaches [21,22] have been proposed to minimizing routing overhead in MANETs. Hybrid routing [16-18] and Connected Dominating Set approaches [21,22] usually reduce routing overhead by solving “rebroadcast redundancy” problem. The key motivation of this paper is to address the both these problems. The central idea is that as long as the above two problems are solved; the route discovery strategy expects that a route with better quality may be found while minimising the routing overhead. The rest of this paper is organized as follows. Details of PRDS are given in Section 2. As an example, a macrobian route strategy based on PRDS is described in Section 3. Simulation results can be found in Section 4. Finally, the paper is concluded in Section 5. II. PRDS In this section, we describe the Priority-based Route Discovery Strategy (PRDS). PRDS uses rebroadcast priority and a competing mechanism to select desirable routes and minimize the routing overhead. Rebroadcast priority is defined using network state parameters. A distributed competing procedure is then executed to determine whether to rebroadcast the RREQ or not. The mechanism of PRDS is quite simple. It assigns a high rebroadcast priority to a “good” candidate for the next hop to solve the “next-hop racing” problem; it uses a competing procedure to prohibit “bad” candidates for the next hop from rebroadcast so as to solve “rebroad-

cast redundancy” problem. In PRDS, a priority index (PI) is defined to indicate how good an intermediate node is for the next hop of the constructing route. PI is defined by some node/link/network state parameters according to different route design purposes such as shortest path, long lifetime path, stable path, load/energy-aware path, and so on. For convenience, we restrict the values of PI from 0 to 1, i.e. PI ∈ [0, 1]. In the next section, we will define a PI for a macrobian route strategy. Then, the RREQ rebroadcast delay is estimated according to PI using the following formula. d = dmax · ( f(PI) + 0.1*random() ) (1) Where, d is the rebroadcast delay of a RREQ; dmax is the upper bound of d. random() is a random function whose values is randomly distributed between 0 and 1. This term is used to differentiate rebroadcast delay when nodes have same PI value. f() is a function of PI that requires the features of (i) a bounded function with upper bound ≤1 and lower bound ≥0; (ii) f() decreases as PI increases. We define the function f() as follows, f(PI) = tanh((1.0-PI)/u0), (2) where tanh(x) is a hyperbolic tangent function; u0 is a constant and the value of 0.3 is appropriate for most cases. f(PI) ∈ [0,0.998] when PI∈ [0,1]. f(PI) decreases rapidly when PI approaches 1 so as to differentiate rebroadcast delay efficiently between high priority nodes. When an intermediate node has determined the rebroadcast delay of a RREQ, it enters a competing procedure to rebroadcast the RREQ. In PRDS, each node needs to maintain a Competing State Table (CST). A CST contains three fields: (i) RREQ ID that is used to identify a unique RREQ. It is represented as (Source ID, broadcast sequence). (ii) The number of times of receiving the same RREQ nh. nh is initialised to 1 when a node receives a first copy of a RREQ. It is also used to represent the competing state. It is set to 0 when the competition is over. Any following RREQs will be deleted as long as their related nh equals 0. (iii) The timestamp of receiving the first copy of the RREQ. This field is used to maintain the CST with soft state mechanism, i.e. timeout mechanism. In PRDS, we define two kinds of events: receiving a RREQ (This event is triggered when a node receives a RREQ.), and a rebroadcast delay time out (This event is triggered when a rebroadcast delay expires.). We pre-assign the number of times (n0) that a RREQ may be received. When a rebroadcast delay times out, PRDS checks the related nh in CST. The node will rebroadcast the RREQ if nh ≤n0. Otherwise, the rebroadcast operation will be prohibited. We denote PRDS using different n0 as PRDS/n0, e.g. PRDS/1, PRDS/2, and so on. The sequence of operations for PRDS is shown in figure 2. As an example to demonstrate its operation, we apply PRDS/1 to the topology in Figure 1. Setting n0 = 1 means that a node will be prohibited from rebroadcasting if it has received more than one copy of the

Waiting for events

no

New?

yes Set nh =1

3 .5

Rebroadcast event triggered

Receiving a RREQ yes

(9 9, 50

no nh++

nh=0? yes

nh=0? or nh >n0

Delete RREQ From buffer

Delete RREQ

Estimate PI, rebroadcast delay; Buffer RREQ; Set a rebroadcast event.

,.1 )

no

Update fields of RREQ Rebroadcast RREQ Set nh =0

Figure 2. PRDS flow chart Figure 3. A route discovery scenario using PRDS/1-MR

RREQ when the rebroadcast delay expires. We simply take DIS/R as PI, where DIS is the distance between the sending and receiving node; R is the transmission range. In figure 1, Node S broadcasts a RREQ that is destined for Node D. Node A, C, and G receive the RREQ and compete for rebroadcast. Node A has the highest rebroadcast priority since Link S-A has the longest length. Node A wins the competition and rebroadcasts the RREQ first. Nodes G and C receive the second copy of the RREQ and thus are prohibited from rebroadcast. Similarly, Node B will rebroadcast the RREQ; Nodes E and H are prohibited from rebroadcast. Note that Nodes F and I will rebroadcast the RREQ because they only receive one copy of the RREQ from Node B (the destination D will not rebroadcast the RREQ). In this example, Node S broadcasts a RREQ; Node A, B, F, and I rebroadcast it in turn; other nodes, i.e. C, E, G, and H are prohibited from rebroadcast. That is, 4/8 of the rebroadcasts are eliminated and the shortest path S-A-B-D is selected. III. PRDS-MR In this section, we design a Macrobian Routing Protocol using PRDS. We term it PRDS-MR. We assume that (i) each node has its own location and mobility knowledge that can be learned from some positioning system; (ii) each node is equipped with an omni-directional antenna that has a transmission range R. PRDS-MR aims at finding the route that has the following features in comparison with RD-random: the lifetime of the route is relatively long; the route length (hops) is not significantly long; routing overhead is minimised. We first define two parameters: Link Alive Time (LAT) and Route Alive Time (RAT). LAT is the amount of remainder time during which two nodes remain connected. RAT is the minimum LAT of the links along the route from source to destination. We denote the coordinates and moving speed of Node i as (xi, yi, zi) and (ui, vi, wi), respectively. The distance between Node 1 and Node 2 can then be expressed as,

DIS = xd2 + y d2 + z d2 ,

(3)

where xd=x1-x2, yd=y1-y2, zd=z1-z2. A link exits between Node 1 and Node 2 if DIS ≤ R, i.e. Node 1 and Node 2 can communicate each other directly. The LAT of the link can be estimated as follows,

LAT =

− ( x d u d + y d v d + z d wd ) + A − B u d2 + v d2 + wd2

,

(4)

where A = (u d2 + v d2 + wd2 ) R 2 ,

B = (ud yd − vd xd ) 2 + (vd z d − wd yd ) 2 + (u d z d − wd xd ) 2 ud=u1-u2, vd=v1-v2, wd=w1-w2 Now, we define PI for PRDS-MR as following, PI = pLAT * pDIS * pRAT, pLAT = tanh((LAT/LAT0)/C1), pDIS = tanh((DIS/R)/C2) , pRAT = tanh((RAT/RAT0)/C3),

(5) (6) (7) (8)

where pLAT is the contribution of LAT of the upstream link. It is the main part of PI. It guarantees that those links with long lifetime have a higher PI. pDIS is the contribution of the length of upstream link. This term prevents very short links from being included in the route. pRAT is the contribution of lifetime of the path from source to the current node. It prevents short lifetime routes from being selected. C1, C2, C3, LAT0, and RAT0 are parameters whose values can be determined by experiment. In this paper, we set these parameters as follows, C1=0.30; C2=0.17; C3=0.05; LAT0=100 sec; RAT0=10 sec. Figure 3 illustrates a route discovery scenario using PRDS-MR. n0 is set to 1 in this scenario. Node S broadcasts a RREQ to discover a route to Node D. The numbers above links are (LAT, RAT, DIS/R); the number

under a link is the PI for the receiving node to rebroadcast the RREQ. For example, numbers (60,50,.9) above the Link A-B mean that LAT of Link A-B is 60 sec; RAT of Route S-A-B is 50 sec; length of Link A-B is 0.9R. The number .96 under Link A-B means the PI for Node B to rebroadcast the RREQ is 0.96. In the figure, Node J is prohibited from broadcast because Link A-J is very short (so pDIS is very small). Node F is prohibited from rebroadcast because the RAT of Path S-E-F is very short (so pRAT is very small). In this example, Path S-A-B-C-D is selected (RAT=50 sec); five nodes are prohibited from rebroadcast. IV. Simulation Results

To evaluate the performance of PRDS-MR, we have implemented PRDS-MR based on AODV. In this section, we conduct simulations in the Global Mobile Simulation (GloMoSim) developing library [23]. We evaluate the performance of PRDS-MR by comparison with AODV. In the simulations, IEEE 802.11 Distributed Coordination Function (DCF) is used as the MAC protocol. The random waypoint model is utilized as the mobility model. In this model, a host randomly selects a destination within the terrain range and moves towards that destination at a speed between the pre-defined minimum and maximum speed. Once the host arrives at the destination, it remains in its current position for a pause time. After the pause time, it randomly selects another destination and speed and then resumes movement. In our simulation, the pause time is set to 0 seconds. The bandwidth of the wireless channel is 2Mbps. The data packet size is 512 bytes. The flow pattern is CBR (Constant Bit Rate). The maximum rebroadcast delay dmax in Eq.1 is set to 30 ms. Table 1 gives the simulation parameters. Table 1. Simulation parameters # node

Area(m2)

50 100 200 500 1000

1500x300 2200x600 3500x900 4500x1500 5500x2500

Simulation time(Sec) 500 500 300 300 300

Other pameters Communication pairs:10; Communication load: 4Pks/source/s; Maximumspeed: 20m/s; Minimumspeed: 0m/s.

The metrics measured in the simulations are: Packet delivery ratio (PDR): the number of packets received by all destinations over the number of packets sent by all sources. Throughput: Size of packets received by all destinations per second (bits/sec). Average end-to-end delay: ∑di/ Nr, where di is the delay experienced by packet i from source to destination; Nr is the number of packets received by all destinations; Routing overhead: number of sending times of routing packets in the simulation. Route lifetime: the interval of a route from its crea-

tion to its broken. While we have conducted extensive simulations, in this paper, we only present a summary of those results that focus on performance and scalability in a dynamic MANET, i.e. how the protocols perform against network size. Protocol scalability is an important issue in MANETs. In some applications such as military communications in a tactical environment, thousands of mobile nodes may exist. In figures 4-9, PRDS/n0-MR denotes simulation results of the protocol using PRDS/n0-MR mechanism, where n0 is the threshold of the RREQ duplicate number as described above. n0=ALL means each node will rebroadcast every non-duplicated RREQ once. Figure 4 illustrates the average route lifetime of each scheme. It shows that route lifetimes of all PRDS/n0-MR schemes are nearly two times as much as AODV. In AODV, RD-random is used in route discoveries; while in PRDS/n0-MR, the long lifetime link has the high priority to be selected. The simulation results demonstrate that PRDS/n0-MR mechanism is effective. Lifetime of the route discovered by PRDS/n0-MR is significant longer than that of AODV. Within PRDS/n0-MR schemes, PRDS/1-MR and PRDS/2-MR perform better than PRDS/ALL-MR. The performance of PRDS against route length is shown in Figure 5. The results show that the average route length of each PRDS/n0-MR scheme is 5~14% longer than that of AODV. This is because PRDS/n0-MR schemes aim at finding the most stable route rather than the shortest route. The macrobian route tends to include shorter links than AODV and leads to longer paths. Considering the benefit from route lifetime, the cost in route length is worthy. PRDS/n0-MR leverages the route length and the route lifetime. Figure 6 shows the same variation tendency of routing overhead for all routing strategies. That is, routing overhead increases as network size increases. However it may be seen that the routing overhead of AODV increases much more rapidly than that of PRDS/n0-MR. The routing overhead of PRDS/ALL-MR, PRDS/1-MR, and PRDS/2-MR is only 42%, 11%, and 15% of that of AODV, respectively, when network size is 1000. The reasons for low routing overhead for PRDS/n0-MR are 1) macrobian route decreases the number of route discoveries; 2) a large amount of rebroadcasts are avoided in the route discoveries. Figure 7 shows the signal collisions measured at MAC layer. IEEE802.11 DCF uses contending-based channel access scheme. It does not have any mechanism to reserve the channel for broadcast. Thus, signal collisions are unavoidable in a real environment. If we compare Figure 7 with Figure 6, we will find that the curves in the two figures are very similar. This implies that signal collisions are highly correlated with routing overhead. Hence, AODV also consumes more bandwidth and energy than PRDS/n0-MR.

20

12 10 8 6 4

AODV PRDS/ALL-MR PRDS/1-MR PRDS/2-MR

16

Route length (hops)

14

Route lifetime (sec)

1000000

18


14 12 10 8 6 4

2

800000


600000

400000

200000

2 0

0

0 0

200

400

600

800

0

1000

200

400

600

800

0

1000

200

Figure 4. Route lifetime

400

600

800

1000

Network size (node number)



Figure 5. Route length

Figure 6. Routing overhead

100

5000000

0.30


90

3000000

Delivery ratio (%)


2000000

80

70

60

1000000


0.25

0.20

Delay (sec)

4000000

Collisions

Routing overhead (packets)

16

0.15

0.10

0.05

50 0.00

0 0

200

400

600

800


Figure 7. Collisions

1000

0

200

400

600

800

1000


Figure 8. Delivery ratio

Figure 8 illustrates the packet delivery ratio. As expected, the packet delivery ratio of each scheme decreases as network size increases. Otherwise, the packet delivery ratio of AODV decreases more quickly than any of the PRDS/n0-MR schemes. The difference between them increases as network size increases. The delivery ratio of PRDS/n0-MR is 2%~5% higher than AODV in 50 node network, while 20%~25% higher in 1000 node network. Among all schemes, PRDS/2-MR performs best; PRDS/1 and PRDS/ALL-MR perform similarly; AODV performs worst. Figure 9 presents the average end-to-end delay of each scheme. In general, delay of any scheme increases as network size increases. In small networks (network size is less than 100), the delay of AODV is lower than that of PRDS/n0-MR. Whereas, the delay of AODV increases rapidly as network size increases and exceeds the delay of PRDS/n0-MR soon. From the above result analyses, we can conclude that PRDS/n0-MR outperforms AODV in terms of packet delivery ratio, end-to-end delay, and route lifetime. At the same time, it decreases routing overhead significantly. It saves much network bandwidth and energy. Other simulations have been made to test adaptivity to dynamics and communication load. These simulation results also indicate that PRDS/n0-MR outperforms AODV.

0

200

400

600

800

1000


Figure 9. Average delay

between rebroadcasting route requests. This has been shown to lead to “next-hop racing” and “rebroadcast redundancy”. In this paper, a novel route discovery strategy termed Priority-based Route Discovery Strategy (PRDS) is described that addresses both these problems. PRDS can be applied to different reactive routing protocols by defining the Priority Index (PI) to find better quality route while decreasing routing overhead. As an example, we have applied PRDS to design a macrobian route discovery strategy (PRDS-MR) and integrate it into AODV. Simulation results show that PRDS-MR outperforms AODV in terms of packet delivery ratio and end-to-end delay while decreasing routing overhead. PRDS-MR has better scalability than AODV. PRDS-MR has additional advantages. It is a distributed algorithm and does not need any periodic messages such as beacons and link state exchanges. A macrobian route requires fewer route reconstructions and hence yields higher throughput and lower routing overhead. A large amount of redundant rebroadcasts are saved and therefore bandwidth and energy consumption by routing overhead is reduced, i.e. improves bandwidth and energy utilization.

Acknowledgement The authors gratefully acknowledge the support and financial assistance provided by the UK EPSRC, under V. Conclusions Current reactive route discovery protocols in Mobile the project GR/S02105/01 “Programmable Routing Ad-hoc Networks (MANETS) employ random delays Strategies for Multi-Hop Wireless Networks”.

References [1] C. E. Perkins and E.M. Royer, “Ad-hoc on demand distance vector routing,” In Proc. WMCSA 1999, New Orleans, LA, pp.90-100. [2] D.B. Johnson and D.A. Maltz, “Dynamic source routing in ad hoc wireless networks,” In Mobile computing, edited by Tomasz Imielinski and Hank Korth, Chapter 5, Kluwer Academic Publishers, 1996, pp.153-181. [3] S. Ni, Y. Tseng, Y. Chen and J. Sheu, “The broadcast storm problem in a mobile ad hoc network,” In Proc. ACM/IEEE MobiCom, Seattle, WA, USA, 1999, pp.151-162. [4] J. Li, and P. Mohapatra, “A novel mechanism for flooding based route discovery in ad hoc networks,” In Proc. IEEE GlobeCom 2003, San Francisco, USA. [5] B. Zhou, J. Wu, X. Fei, J. Zhao, “PCBA: A priority-based competitive broadcasting algorithm in mobile ad hoc networks,” J. Computer Sci. & Technol., 2003.18(5), pp.598-606. [6] C.K. Toh, “Associativity based routing for ad hoc mobile networks,” Wireless Personal Communications Journal, Special Issue on Mobile Networking and Computing Systems, 4(2), March 1997, pp.103–139. [7] R. Dube et al, “Signal stability based adaptive routing (SSA) for ad-hoc mobile networks,” IEEE Pers. Commun., Feb. 1997, pp. 36–45. [8] S.J. Lee, M. Gerla, “Dynamic load-aware routing in ad hoc networks,” In Proc. of IEEE ICC 2001, Helsinki, Finland, June 2001, pp. 3206-3210. [9] H. Hassanein and A. Zhou, “Routing with load balancing in wireless ad hoc networks”, In Proc. ACM MSWiM 2001, Rome, Italy, pp.89-96. [10] K. Wu and J. Harms, “Load-sensitive routing for mobile ad hoc networks,” in Proc. IEEE ICCCN 2001, Scottsdale, AZ. [11] V. Rodoplu and T.H Meng, “Minimum energy mobile wireless networks,” IEEE Journal on Selected Areas in Communications, Vol. 17, No.8, 1999, pp.1333-13. [12] S. Singh, M. Woo and C.S. Raghavendra, “Power-aware routing in mobile ad hoc networks,” In Proc. ACM/IEEE MOBICOM 1998, pp.181-190. [13] M.W. Subbarao, “Dynamic power-conscious routing for MANETs: an initial approach,” In Proc. IEEE VTC 1999, Amsterdam, Netherlands, pp.1232-1237. [14] R. Castaneda and S. Das, “Query localization techniques for on-demand routing protocols in ad hoc networks,” In Proc. ACM/IEEE MOBICOM 1999, pp.186-194. [15] N. Kothari, V. Bhandari and D. Sanghi, “Query localization using pheromone trails: A swarm intelligence based approach,” In Proc. NCC'03, Chennai, India. [16] M. Jiang, J Li, and Y.C.Yay, “Cluster based routing protocol (CBRP),” Internet-Draft, draft-ietf-manet-cbrp-spec-01.txt, August 1999. [17] Z.J.Hass, “A new routing protocol for the reconfigurable wireless networks,” In Proc. ICUPC’97, San Diego, CA USA, pp.562-566. [18] M. Joa-Ng and I.T. Lu, “A peer-to-peer zone-based two-level link state routing for mobile ad hoc networks,” IEEE Journal on Selected Areas in Communications, Special Issue on Ad-Hoc Networks, Aug. 1999, pp.1415-25. [19] J. Li and P. Mohapatra, “LAKER: Location aided knowledge extraction routing for mobile ad hoc networks,” In Proc. WCNC 2003, New Orleans, Louisiana,USA. [20] Y.B. Ko and N.H. Vaidya, “Location-aided routing (LAR) in mobile Ad hoc networks,” In Proc. ACM/IEEE MOBICOM 1998, Dallas, TX, USA, pp.66-75. [21] J. Wu, M. Gao and I. Stojmenovic, “On Calculating Power-Aware Connected Dominating Sets for Efficient Routing in Ad Hoc Wireless Networks,” In Proc. ICPP 2001, Valencia, Spain. [22] B. Das and V. Bharghavan, “Routing in ad-hoc networks using minimum connected dominating sets,” Communications, 1997. ICC ’97 Montreal, Towards the Knowledge Millennium. 1997 IEEE International Conference on, Vol. 1, 1997, pp. 376-380. [23] L. Bajaj, M. Takai, R. Ahuja, K. Tang, R. Bagrodia, and M. Gerla, "GloMoSim: A scalable network simulation environment," UCLA Computer Science Department Technical Report 990027, May 1999.