Upload
finn-gardner
View
30
Download
2
Embed Size (px)
DESCRIPTION
Joint Design of Routing and Medium Access Control for Hybrid Mobile Ad Hoc Networks. Mobile Networks and Applications (January 2007) Presented by J.H. Su ( 蘇至浩 ). Authors(1/2). Xiaojiang (James) Du Received the Ph.D. degrees from University of Maryland College Park in electrical engineering. - PowerPoint PPT Presentation
Citation preview
Mobile Networks and Applications (January 2007)
Presented by J.H. Su (蘇至浩 )
112/04/19OPLab, IM, NTU1
Joint Design of Routing and Medium Access Control
for Hybrid Mobile Ad Hoc Networks
Authors(1/2)
112/04/19OPLab, IM, NTU2
Xiaojiang (James) DuReceived the Ph.D. degrees from University of
Maryland College Park in electrical engineering.
An Assistant Professor with the Department of Computer Science, North Dakota State University, Fargo.
His research interests are wireless sensor networks, mobile ad hoc networks, wireless networks, computer networks, network security, and network management.
Authors(2/2)
112/04/19OPLab, IM, NTU3
Dapeng WuReceived the Ph.D. degree in electrical and computer
engineering from Carnegie Mellon University, Pittsburgh, PA, in 2003.
Has been with Electrical and Computer Engineering Department, University of Florida, Gainesville, as an Assistant Professor.
His research interests are in the areas of networking, communications, multimedia, signal processing, and information and network security.
Agenda
112/04/19OPLab, IM, NTU4
Introduction
Assumption of HR
Hybrid Routing Protocol
Performance Evaluation
Conclusion
Agenda
112/04/19OPLab, IM, NTU5
Introduction
Assumption of HR
Hybrid Routing Protocol
Performance Evaluation
Conclusion
Introduction (1/2)
112/04/19OPLab, IM, NTU6
Most existing routing protocols assume homogeneous MANETs.
Gupta and Kumar showed that per node capacity of a homogeneous wireless network is only w/√ (nlogn)W is the node transmission capacity n is the number of nodes
Introduction (2/2)
112/04/19OPLab, IM, NTU7
Furthermore, in many realistic ad hoc networks, nodes are hybrid.in a battlefield network
In this paper, we present a new routing protocol—Hybrid Routing (HR) protocol for hybrid MANETs.node location information is used to reduce
routing overhead
Agenda
112/04/19OPLab, IM, NTU8
Introduction
Assumption of HR
Hybrid Routing Protocol
Performance Evaluation
Conclusion
Assumption of HR (1/2)
112/04/19OPLab, IM, NTU9
For simplicity, we consider there are only two types of nodes in the network.Backbone-Capable nodesgeneral nodes
The main idea of HR protocol is to let most routing activities rely on B-nodes.provides better reliability and fault tolerancerouting packets via B-nodes is more efficient
than using general nodesreduces the routing overhead and latency
Assumption of HR (2/2)
112/04/19OPLab, IM, NTU10
The routing area is divided into several small, equal-sized squares—referred to as cells.
In HR protocol, we assume the routing area is fixed. the position of each cell is also fixed
There is a unique id for each cell.
One (and only one) B-node is elected and maintained in each cell, and each B-node has a second address.
B-node can always communicate directly with B-nodes in all nearby cells.
Agenda
112/04/19OPLab, IM, NTU11
IntroductionAssumption of HRHybrid Routing
Routing among B-nodesRoute DiscoveryRoute RepairDissemination of Node Location InformationElection of B-nodeHybrid MAC (HMAC)
Performance EvaluationConclusion
Routing among B-nodes (1/2)
112/04/19OPLab, IM, NTU12
B-nodes use their second addresses to communicate with each other.
A straight line L is drawn between the centers of cell CS and Cd.
Two border lines (blue lines) which parallel to line L with distance of W from L are drawn from CS to Cd.the value of W depends on the density of BC-nodes in
the network
All the cells that are within the two border lines are defined as routing cells.
Routing among B-nodes (2/2)
112/04/19OPLab, IM, NTU13
W
W
Agenda
112/04/19OPLab, IM, NTU14
IntroductionAssumption of HRHybrid Routing
Routing among B-nodesRoute DiscoveryRoute RepairDissemination of Node Location InformationElection of B-nodeHybrid MAC (HMAC)
Performance EvaluationConclusion
Route Discovery (1/4)
112/04/19OPLab, IM, NTU15
Consider a node S wants to send a data packet to a destination node D.Case 1 : S is a B-nodeCase 2 : S is a general node
S needs to know the current location of the destination node D.
Then S determines the routing cells between S and Bd, and sends Route Request (RR) packets to B-nodes in routing cells.
Route Discovery (2/4)
112/04/19OPLab, IM, NTU16
The RRpacket includes the following fields:Starting_B-nodeNext_cells Routing_cells Path Destination_cell
Route Discovery (3/4)
112/04/19OPLab, IM, NTU171 2 3 4
5 6 7 8
9 10 12
13 14 16
11
S
D
Route Discovery (4/4)
112/04/19OPLab, IM, NTU18
If S does not receive a RP from Bs for a certain time:S will initiate a B-node elecion
If Bs does not receive a RP from Bd for a certain time:Bs will flood the RR packet to all B-nodes in the network
If Bd does not receive the Ack for a certain timeBd will hold the data packet (before receiving Ack) and
request the updated location information of node Dthen Bd can send the data packet to the B-node
closest to the new location of D
Agenda
112/04/19OPLab, IM, NTU19
IntroductionAssumption of HRHybrid Routing
Routing among B-nodesRoute DiscoveryRoute RepairDissemination of Node Location InformationElection of B-nodeHybrid MAC (HMAC)
Performance EvaluationConclusion
Route Repair (1/2)
112/04/19OPLab, IM, NTU20
Route Repair (2/2)
112/04/19OPLab, IM, NTU21
The RE packet includes the following fields:Repairing_B-nodeNext_cellRouting_cellsDestination_cellDestination_ nodePath
Agenda
112/04/19OPLab, IM, NTU22
IntroductionAssumption of HRHybrid Routing
Routing among B-nodesRoute DiscoveryRoute RepairDissemination of Node Location InformationElection of B-nodeHybrid MAC (HMAC)
Performance EvaluationConclusion
Dissemination of Node Location Information
112/04/19OPLab, IM, NTU23
When a node moves out of its previous cell, it sends a location update packet (with its new location) to the B-node in the new cell (or the nearest B-node).
All B-nodes periodically send aggregated node location information to a special B-node B0.
If B0 also moves around, then B0 needs to multicast its current location to all B-nodes
Agenda
112/04/19OPLab, IM, NTU24
IntroductionAssumption of HRHybrid Routing
Routing among B-nodesRoute DiscoveryRoute RepairDissemination of Node Location InformationElection of B-nodeHybrid MAC (HMAC)
Performance EvaluationConclusion
Election of B-node (1/2)
112/04/19OPLab, IM, NTU25
Initially, one B-node is elected in each cell if there are BC-nodes available in the cell.
Following events Initiates the B-node election process:a B-node moves out of its current cella general node discovers there is no B-node
in the cell
Election of B-node (2/2)
112/04/19OPLab, IM, NTU26
The election process works as following:the leaving B-node or the general node floods
an election message to all the nodes in the cell
when a BC-node receives the election message, it broadcasts a claim message that claims it will become the B-node to all nodes in the cell (each BC-node defers a random time before its B-node claim)
if other BC-node hears a claim message during this random time, it then gives up its broadcast.
Agenda
112/04/19OPLab, IM, NTU27
IntroductionAssumption of HRHybrid Routing
Routing among B-nodesRoute DiscoveryRoute RepairDissemination of Node Location InformationElection of B-nodeHybrid MAC (HMAC)
Performance EvaluationConclusion
Hybrid MAC (1/2)
112/04/19OPLab, IM, NTU28
The key idea is to combine time-slotted mechanism with contention based mechanism.
A large time frame is divided into three sub-framesG-to-GB-to-BB-to-G
Hybrid MAC (2/2)
112/04/19OPLab, IM, NTU29
Inside each sub-frame, the corresponding nodes use contention-based mechanism— IEEE 802.11b to decide which node should transmit packets.
The sub-frame length depends on the amount of traffic of each subtype.
Agenda
112/04/19OPLab, IM, NTU30
Introduction Assumption of HR Hybrid Routing Protocol Performance Evaluation
Experiment Setting Routing overhead under different mobility Routing overhead vs. transmission range Throughput under different traffic load Delay comparison Scalability The performance of HMAC The probability of having B-nodes Performance for different BC-node densities
Conclusion
Experiment Setting(1/3)
112/04/19OPLab, IM, NTU31
there are totally 20 time-slotsG-to-G sub-frame takes eight timeslotsB-to-B sub-frame takes eight time-slotsB-to-G sub-frame takes four time-slots
we present the simulations performed by distributing 100 nodes uniformly at random in an area of 1,000×1,000 m.25 Backbone-Capable (BC) nodes75 General nodes
Experiment Setting(2/3)
112/04/19OPLab, IM, NTU32
The transmission range of a B-node and a general node is 400 and 100 m.
The side length of a cell is set as a=R/1.6.
There are 16 cells in the routing area
The width of the routing cells—W is set to 0
Experiment Setting(3/3)
112/04/19OPLab, IM, NTU33
Each simulation was run for 600 simulated seconds.
The source sent packet of 512 b at a rate of four packets per second.
We ran each simulation ten times to get an average result for each simulation configuration.
We compared our HR protocol with AODV (Ad-hoc on demand distance vector) routing protocol.
Agenda
112/04/19OPLab, IM, NTU34
Introduction Assumption of HR Hybrid Routing Protocol Performance Evaluation
Experiment Setting Routing overhead under different mobility Routing overhead vs. transmission range Throughput under different traffic load Delay comparison Scalability The performance of HMAC The probability of having B-nodes Performance for different BC-node densities
Conclusion
Routing overhead under different mobility
112/04/19OPLab, IM, NTU35
Agenda
112/04/19OPLab, IM, NTU36
Introduction Assumption of HR Hybrid Routing Protocol Performance Evaluation
Experiment Setting Routing overhead under different mobility Routing overhead vs. transmission range Throughput under different traffic load Delay comparison Scalability The performance of HMAC The probability of having B-nodes Performance for different BC-node densities
Conclusion
Routing overhead vs. transmission range
112/04/19OPLab, IM, NTU37
Agenda
112/04/19OPLab, IM, NTU38
Introduction Assumption of HR Hybrid Routing Protocol Performance Evaluation
Experiment Setting Routing overhead under different mobility Routing overhead vs. transmission range Throughput under different traffic load Delay comparison Scalability The performance of HMAC The probability of having B-nodes Performance for different BC-node densities
Conclusion
Throughput under different traffic load
112/04/19OPLab, IM, NTU39
Agenda
112/04/19OPLab, IM, NTU40
Introduction Assumption of HR Hybrid Routing Protocol Performance Evaluation
Experiment Setting Routing overhead under different mobility Routing overhead vs. transmission range Throughput under different traffic load Delay comparison Scalability The performance of HMAC The probability of having B-nodes Performance for different BC-node densities
Conclusion
Delay comparison
112/04/19OPLab, IM, NTU41
Agenda
112/04/19OPLab, IM, NTU42
Introduction Assumption of HR Hybrid Routing Protocol Performance Evaluation
Experiment Setting Routing overhead under different mobility Routing overhead vs. transmission range Throughput under different traffic load Delay comparison Scalability The performance of HMAC The probability of having B-nodes Performance for different BC-node densities
Conclusion
Scalability
112/04/19OPLab, IM, NTU43
HR-S HR-L
AODV-S
AODV-L
Agenda
112/04/19OPLab, IM, NTU44
Introduction Assumption of HR Hybrid Routing Protocol Performance Evaluation
Experiment Setting Routing overhead under different mobility Routing overhead vs. transmission range Throughput under different traffic load Delay comparison Scalability The performance of HMAC The probability of having B-nodes Performance for different BC-node densities
Conclusion
The performance of HMAC- Throughput
112/04/19OPLab, IM, NTU45
The performance of HMAC- Delay
112/04/19OPLab, IM, NTU46
Agenda
112/04/19OPLab, IM, NTU47
Introduction Assumption of HR Hybrid Routing Protocol Performance Evaluation
Experiment Setting Routing overhead under different mobility Routing overhead vs. transmission range Throughput under different traffic load Delay comparison Scalability The performance of HMAC The probability of having B-nodes Performance for different BC-node densities
Conclusion
Performance for different BC-node densities
112/04/19OPLab, IM, NTU48
Agenda
112/04/19OPLab, IM, NTU49
Introduction
Assumption of HR
Hybrid Routing Protocol
Performance Evaluation
Conclusion
Conclusion
112/04/19OPLab, IM, NTU50
Proposed Hybrid Routing Protocol to resolve routing problem in different type of node.
An efficient algorithm is provided to disseminate node location information among all B-nodes.
Proposed a novel HMAC protocol that can improve the efficiency of medium access in hybrid MANETs.
Conclusion
112/04/19OPLab, IM, NTU51
HR has very good performance, and performs much better than AODV in terms of reliability, scalability, route discovery latency, routing overhead as well as packet delay and throughput.