9 ad hoc network protocols

The Graduate School of Information Technology and TelecommunicatThe Graduate School of Information Technology and Telecommunications, INHA Universityions, INHA Universityhttp://http://multinet.inha.ac.krmultinet.inha.ac.kr

Multimedia Network Lab.Multimedia Network Lab.

Mobile Computing

Chapter 9: Ad-hoc Protocols

Prof. Sang-Jo Yoo

2The Graduate School of Information Technology and TelecommunicatThe Graduate School of Information Technology and Telecommunications, INHA Universityions, INHA University

http://multinet.inha.ac.kr Multimedia Network LabMultimedia Network Lab..

ContentsIntroduction

Ad-hoc Network Overview

Routing over Multi-hop Ad-hoc Networks

DSDV Protocol

DSR Protocol

AODV Protocol

ZRP Protocol

Conclusion Remarks



Introduction



Model of OperationWireless computing devices should physically be able to communicate with each other, even when no routers or base stations or Internet service providers (ISPs) can be found.

In the absence of infrastructure, what is needed is that the wireless devices themselves take on the missing functions.



Model of OperationAd-hoc network model

The nodes are using IP and they have IP addresses.The nodes are far enough apart so that not all of them are within range of each other.The nodes may be mobile so that two nodes within range at one point in time may be out of range moments later.The nodes are able to assist each other in the process of delivering packets of data.

MH3

MH1

MH4

MH2

MH1

MH6

MH5

MH7

MH8



ApplicationsConferencing

Sometimes business network infrastructure is often missing (e.g., outside of their normal office), but the need of collaborative computing might be even more important.

Home NetworkingOffers the prospect of reachability to all the nodes at home regardless of their “normal” point of attachment.

Emergency ServicesAs the Internet (existing infrastructure) grows in importance, the loss of network connectivity during natural disasters will become an ever more noticeable consequence of the calamity.Very important to fine ways to enable the operations of networks even when infrastructure elements have been disabled.



ApplicationsAccidents or disasters in rural areas where poor or no preestablished communication infrastructure exists:

Extending infrastructure network coverageRapid deployment of communication infrastructure in rural areas.

InfrastructureNetwork

Gateway

Ad Hoc Network



ApplicationsPersonal Area Networks and Bluetooth

PAN: very localized network populated by some network nodes that are closely associated with a single person.The PAN devices may not need to have an attachment to the public networks, but certainly need to communicate with each other.Intra PAN communication and inter PAN communication

Ubiquitous ComputingComputers will be all around us, constantly performing mundane tasks to make our lives a little easier.

전략적 중요도

매우 낮음 낮음 보통 높음 매우 높음

매우낮음

낮음

보통

높음

매우높음

경제적파급효과

유비퀴터스 네트워크(2005년 2조300억 달러)

이동 및 착용형정보통신기기

(2005년 733억 달러)

전략적 중요도

매우 낮음 낮음 보통 높음 매우 높음

매우낮음

낮음

보통

높음

매우높음

경제적파급효과

유비퀴터스 네트워크(2005년 2조300억 달러)

이동 및 착용형정보통신기기

(2005년 733억 달러)



Applications“Ubiquitous intelligent internetworking devices that detect theirenvironment, interact with each other, and respond to changing environmental condition will create a future that is as challenging to imagine as a science fiction scenario.”

Sensor DustA large collection of tiny sensor devices

once situated, the sensors remain stationarylargely homogeneouspower is likely to be a scarce resource, which determines the lifetime of the network

Aan offer detailed information about terrain or environmental dangerous conditions.



ApplicationsSensor Networks

Distribute sensors containing wireless transceivers.The sensors could then form an ad-hoc network and cooperate to gather the desired information.

Transceiver

Memory

SensorsBattery

EmbeddedProcessor

1kbps-1Mbps10-100 Meters128KB-1MB

Limited Storage

60% of total cost Limited

life time

10MHzLow costprocessor

Transceiver

Memory

SensorsBattery

EmbeddedProcessor

Transceiver

Memory

SensorsBattery

EmbeddedProcessor

1kbps-1Mbps10-100 Meters128KB-1MB

Limited Storage

60% of total cost Limited

life time

10MHzLow costprocessor



ApplicationsAutomotive/PC Interaction

Ad-hoc network between automotive computers and laptops or PDA in our cars.

Indicate mechanical problems, repair service categories, repair shop position, and so on.

Ad-hoc network between automobiles and their occupants.Positioning information, …

Ad-hoc network between cars.

Other Envisioned Applications



Technical and Market FactorsScalability

Node mobility: routing information is changed very rapid.Ad-hoc network do not typically allow the same kinds of aggregationtechniques for standard Internet routing protocols.Scalability problemNeed some ways to maintain aggregation for ad-hoc networks.Ad-hoc network should be designed to reduce the number of control messages.Ad-hoc network algorithms should be carefully evaluated and compared for their relative scalability in the face of node population growth and increased node mobility.

Power Budget vs. Latency



Ad-hoc Network Overview



MANETMobile Ad-hoc Network (MANET)

A collection of mobile platforms nodes where each node is free to move about arbitrarily. Distributed, mobile, wireless, multi-hop networks

Without the benefit of any existing infrastructure.Each node logically consists of a router.



MANETBackground

Since 1970s, U.S. DoD sponsored research on this network.

MANET can expand the Internet.



Characteristics of MANETDynamic topology

links formed and broken with mobility

Possibly uni-directional linksConstrained resources

battery powerwireless transmitter range

Network partitions

A B AB



MANET Design IssuesNetwork size

Node density & geographical area

ConnectivityThe number of neighbors that each node can link to directly.Link capacity between nodes.

Network topologyUser mobility can directly affect on the network topology.

User trafficTraffic type: burst or constantTraffic priority and MAC layer issue

Operational environmentEnergy

Battery-operated store-and-forward nodes Need low energy networking approachLink, MAC, Network layer design issues



MANET Design IssuesRegulatory

Power spectral density (PSD) requirement for both legal and public health reasons.

Performance metricsQoS requirements

Throughput, delay, loss

CostCost-versus-performance trades



HistoryOriginal motivation for MANET

Military need for battle field survivability.Mobile wireless communication systems for coordinating group actions in a distributed manner, avoiding single points of failure.

Military need to set up a temporal network when there is no fixed and pre-placed communications infrastructure.Need for multi-hop networking

Low power, LOS (line of sight) limitation, and so on…



HistoryPacket Radio Network (PRNet)



Nokia Rooftop Product



FHPFHP Wireless, USA

ad hoc network in a campus



FHP Wireless



MeshNetworksMeshNetworks, USA



Networking Scenario: To Internet



SkyPilot NeighborNetSkyPilot Network, USA



Routing over Multi-hop Wireless

Networks



Review of RoutingNext-hop routing:

Each node maintains for each destination the preferred neighbor (i.e., next hop)

Source routing:The source keeps for each destination the whole path leading to each destination.



Link-State Routing:Each node periodically broadcasts the link costs of its outgoing links to all destinations.

could be done through floodingAs a node receives this information, it updates its view of the network topology.

Each node knows the entire network topology.Each node determines the best path to the destination.

ex: based on shortest-path algorithm

...

5cy

3ax

costnext hopdestination

Review of Routing



Distance-Vector Routing:Host i, for each destination x, keeps a set of distances {dij(x)| j is a neighbor or i}.

Host k is the next hop if dik(x) = min{dij(x), for all j’s}.Update Rule:

Each node monitors the costs of its outgoing links and periodically broadcasts, to all neighbors, its current estimate of the shortest path to every other node.

...

5j2

3j1

distance dij(x)neighbor

Review of Routing



Ad-hoc Routing OverviewTable Driven (Proactive) Protocol

When the application starts, a route can be immediately selected from the route table

Sore route information even before it is neededTraditional distributed shortest-path protocolsMaintain routes between every host pair at all times

Additional control packets: continually update stale route entries.Example: DSDV (destination sequenced distance vector)



Ad-hoc Routing OverviewOn Demand (Reactive) Protocol

Routing information is acquired only when it is actually needed.Less bandwidth for maintain the route table.

Source initiates route discoverySuffer a long delay when the application starts

Because a route to the destination will have to be acquired before the communications can begin.

Example: DSR (dynamic source routing)

Hybrid ProtocolAdaptive; Combination of proactive and reactiveExample : ZRP (zone routing protocol)



Ad-hoc Routing Overview

Ad hoc routing protocol

Proactive(Table-driven) Reactive(On-demand)Hybrid

DSDV WRP

CGSR

ZRP AODV DSR LMR ABR

TORA SSR



DSDV Protocol(Destination Sequenced Distance

Vector)



DSDV OverviewDSDV: Charles E. Perkins & Pravin Bhagwat, Computer

Communications Review, ’94

Routing table follows the distance-vector style.periodical update

Each route entry is tagged with a sequence number that is originated by the destination node.

to avoid loops

PurposeConstruction of temporary networks with no wires and no administrative intervention required.

Previous routing protocols for wired networks to ad-hocHighly dynamic network topologyToo heavy (computational burden)Convergence characteristics do not seem good enough Wireless medium differs



Route AdvertisementEach MN advertises its own route table to each of the current neighbors.

The entries in this list may change dynamically over time.A lot of control traffic in the network.

Solution two types of route update packets Full dump

All available routing information

Incremental Only information changed since the last full dump

Use hop count as the metric.could be easily extended to other metrics.

Route Advertisement (DSDV)



Route Advertisement (DSDV)Special designs:

It sometimes delays advertising routes to damp fluctuations of routing tables.

to avoid broadcasting unstable route informationSometimes speedup advertising routes to propagate important topology changes.

such as new links, broken links.

Each node periodically broadcasts its route entries.say, every few seconds“broadcast” is limited to a physical range (due to the nature of wireless communication)

not network-wideunreliable

power-saving hosts are not disturbed.



Route Table (DSDV)The information in each broadcast:

Route entry:destination’s addresshop count to the destinationsequence number (originally “stamped” by the destination node)

to distinguish fresh from stale route entrieseach broadcast has a new number (typically, +1)

Within the header of the packet route table also containssender’s new sequence numberhardware address (MAC) and network address (IP)

for address resolution from layer-3 to layer-2

Route SelectionMore recent seq num is preferred.Paths with the same seq num those with the smallest metric will be used.



DSDV Operation (DSDV)

MH3

MH1

MH7

MH8

MH5

MH6

MH4

MH1

MH2

The install time helps determine

when delete stale routes.



DSDV Operation (DSDV)



DSDV Operation (DSDV)Procedure

When MH1 moves into the vicinity of MH8 and MH7, it triggers an immediate incremental routing information update, which is broadcast to MH6.MH6 also triggers an immediate update, which carries along the new routing information for MH1.MH4, upon receiving this information, then broadcasts it at every interval until the next full routing information dump.At MH4, the incremental advertised routing updates has the form shown in Table 4.

Quick Update:When a new or substantially modified route is received, it is broadcast immediately.

Periodical Update:Otherwise, route entries are rebroadcast periodically.



Quick Update (DSDV)When a Quick Update Is Needed?

Case 1: when X finds its (direct) next hop Y is brokenX assigns an infinity metric to host Y.X increases the seq. no. of Y by 1.

This is the ONLY situation that Y’s destination sequence number can be increased by a host other than Y.

X broadcasts the route info. immediately.

Case 2: when X knows a route to a distant host Z and receives an infinity metric for Z with an old or equal sequence no:

broadcast its own route information immediately



X Y

dest_seq_no of Y = k

route entry for Y:dest_seq_no of Y = kmetric = 1

broadcast:dest_seq_no of Y = k+1metric = infinity

Quick Update: case 1 (DSDV)



Quick Update: case 2 (DSDV)

X Y

step 1:Y: “I have no route to Z.”

(seq_no(Z) = 10)

step 2: (X’ broadcast) “I can reach Z with 6 hops.”(seq_no = 11)

Z

route entry for Z:seq_no(Z) = 11metric = 6



Damping Fluctuations (DSDV)How the settling time table is used to prevent fluctuations of route table entry advertisement.

MH9

MH2 MH6

MH4

Mobile HostCollection 2

Mobile HostCollection 1

The larger the number of the updates can be expected in Figure

The settling time data is stored in a table with the following fields.

1) Destination address

2) Last settling time

3) Average settling time 1. 12 hops(SN=10)

3. 12 hops(SN=11)………..

2. 11 hops(SN=10)

4. 11 hops(SN=11)

……...



Damping Fluctuations (DSDV)Suppose a new routing information update arrives at MH4 and the SN in the new entry is newer than the SN in the currently used entry but has a worse metric,

Then MH4 must use the new entry in making subsequent forwarding decisions.MH4 does not have to advertise the new route immediately and canconsult its route settling time table to decide how long to wait before advertise it.The average settling time is used for this determination. For instance, MH4 may decide to delay (average_settling_time*2) before advertising a route.

The settling time is calculated by maintaining a running , weighted average over the most recent updates of the routers, for each destination.

This can be quite beneficial, because if the possibly unstable route were advertised immediately, the effects would ripple through the network.



DSDV design goalThe DSDV approach relies on periodic exchange of routing information among all participating nodes.

An alternative is to design a system that performs route discovery on a need to know basis. For devices operating on limited battery power, this may be an important design consideration.



DSR Protocol(Dynamic Source Routing Protocol)



What is DSR?DSR is self-organizing and self-configuringComposed of two mechanisms

Route DiscoveryRoute Maintenance

Source Routing- the sender of a packet determines the complete sequence of nodes to forward the packet

Designed for use in wireless ad hoc networksNo periodic router advertisements Dynamically determines route based on cached info and on the results of a route discovery



TermsFound on Packets

Source RoutesThe path of nodes the packet will travel.

Route RecordAddress of each intermediate node through the Route Request message has been forwarded. (initialized to an empty list)

Request IDUnique identifier determined by the initiator

Found in Nodes

Send BufferContains a copy of each pack that cannot be transmitted because of no source route

Route CachePreviously learned source routes



Goal is to find a Source RouteNormally will find the source route by searching its Route CacheIf no route is found then the initiator will send a Route Request

A host initiating a route discovery broadcasts a Route Request packetEach Route Request packet contains a route record, request idIf successful, initiating host receives a Route Reply packet

The route is saved in the cache for future use.

Basic Operations in DSR: Route Discovery



Example: Route DiscoveryA sends the route request packet to find the route to H

A

B

D

GE

F

C H

Building Record Route during Route Discovery

source

destinationID=2

<A>

<A>

<A>ID=2

ID=2

<A,D>

<A,B>

<A,C>

ID=2

ID=2

ID=2 <A,C,E>

<A,D,F>ID=2

ID=2<A,C,E,G>

ID=2



When Route Request Receivedif (node == target)

send Route Reply to initiator;

if (node has already seen the Route Request with the same ID)discard Route Request;

else append address to route record and broadcast;

When the destination receives the Route Request message.Will send a Route Reply messageH will examine it’s Route Cache for a route back to node A.If no path found, H will send its own Route Request for node A with Route Reply piggybacked.

Unidirectional link

Symmetric link: the reverse direction exists.



Example: Route Discovery

Propagation of Route Reply with the Route Record

A

B

D

GE

F

C H

<A,D,F>

<A,D,F>

<A,D,F>ID=2

ID=2

ID=2

If the source node attempts to send additional data packets to this same node. In fact, it should initiate Route Request because it is possible that the destination node is not currently reachable.To reduce the overhead, If the node attempts to send additional data packets to this same node more frequently than the limit allows, the subsequent packets should be buffered in the Send Buffer until a Route Reply is received, but if the node does send packets within the minimum allowable interval between new Route Discoveries for the same target, then the node can send all packets without initiating any Route Requestmessage.



Route Maintenance

If transmission error is detected at data link layer, Route Errorpacket is generated and sent to the original sender of the packet.

The node removes the broken link in error from its route cache when a Route Error packet is received.

Retransmission is a function for upper layer (TCP).To destination EIf A has in its Route Cache another route to E, it sends the packet using the new route immediately.Otherwise, it may perform a new Route Discovery for E.

A B C D E



Additional Route Discovery FeaturesCaching Overheard Routing Information

A node forwarding or overhearing any packet may add the routing information from that packet to its own Route Cache.

From data packet, Route Request, and Route Reply.Limitation on caching

When unidirectional links are used.C forwards a data packet along the route from A to E.It learns from the header, C D E possible route.But C B A may not work because these links might be unidirectional.

A B C D E



Additional Route Discovery FeaturesReplying to Route Requests Using Cached Routes

A node (not target node) receiving a Route Request searches its own Route Cache.If a route is found the node generally returns a Route Reply to the initiator itself rather than forwarding the Route Request.

<A Route Reply Storm>Nodes B,C,D,E, and F all receive A’s Route Request for target G.



Additional Route Discovery FeaturesPreventing Route Reply Storms

The above replying functionality can result in a possible Route Reply“storm”.

If a node broadcasts a Route Request for a target node.Wasting bandwidth and increasing network collisions.

Delay sending its own Route Reply for a short period.D=H * (h-1+r)

H=constant delay, h=hop numbers to the route, r=random number between 0 and 1.

Route Request Hop LimitsTo limit the number of intermediate nodes allowed to forward theRoute Request.As the Request if forwarded, the limit is decreased.

If it reaches zero before the target, it is discarded.



AODV Protocol(Ad Hoc On-Demand Distance-

Vector Protocol)



Design Choices

DSDVPro-active

routing table entry for everydestinationperiodically updateShort establishment delayOverhead for unnecessary link information (O(n2))

• Hop-by-hop routingLess per-packet overhead

DSRReactive

route for communicated destinationson-demand routing: floodingLong routing delay (reduced by cache/overhearing)Less overhead packet

Source routingLarge per-packet overheadLess scalable in term of #hop

AODV = Hop-by-hop + Reactive



AODV Design PurposesAd-hoc On-Demand Distance Vector Routing, Charles E. Perkins, Elizabeth M. Royer. Proceedings of IEEE WMCSA'99,New Orleans, LA, Feb. 1999

Quick adaptation under dynamic link conditionsLower transmission latencyLow network utilization (less broadcast)Loop-free property (How: using destination sequence #)Scalable to large network

AssumptionsUses symmetric links.



Key Ideas“Partial” routing table are constructed lazily

Entries are updated only when a node sends to another node unavailable in its routing tableNo periodic updatesNode not on active paths maintain no routing entriesReduce packet overhead

Routing tableNo source routing needed reduce bit overheadServe as “cache” reduce establishment latencySequence number loop free

Push link failure to relevant nodesReduce establishment latency



Route Establishment ProcedureRoute Discovery: on demand

Follows route request (RREQ) and route reply (RREP) messages.

1. When, a node needs a route to a destination it broadcasts a RREQ.

2. Any node with a route to the destination can reply RREP to the source node.

3. Route information is maintained by each node in its route table.

4. Information obtained through RREQ and RREP is kept in the route table.

5. Sequence numbers are used to eliminate stale routes.

6. Routes with old sequence numbers are aged out of the system.



Route DiscoveryInitiated when a node needs to communicate with another node which has no information in the local route table.

Each node has 2 counters:

Source node broadcasts a route request (RREQ) to its neighbor nodes.

broadcast_idnode sequence number

hop_cnt

dest_addr

source_addr

source_sequence_#

dest_Sequence_#

broadcast_id



RREQ - details<source_addr, broadcast_id> is unique

Broadcast_id is incremented for new RREQ

If the neighboring node doesn’t reply with a RREP, hop_cnt is incremented.

RREQ from same node with same broadcast_id will not be broadcasted more than once.

AB

C

D

B won’t rebroadcast this RREQ

For example, node A wants to contact node D, but there is no active path to reach node D.



RREQ - detailsSource_sequence_#: To maintain freshness information about the reverse route to the source

Dest_sequence_#: how fresh a route to the destination

Every node will record the neighbor’s address where first copy of RREQ is from

These entries will be maintained for long enough for RREQ to traverse and produce a RREP to the sender



Process the RREQTo process the RREQ, the node sets up a reverse route (path) entryfor the source node in its route table.

Source node’s IP addressSequence numberNumber of hops to the source nodeIP address of the neighbor from which the RREQ was received – the next hop toward the source node.

With reverse route entry, the node knows how to forward a RREP to the source.

If the route entry is not used within the specified lifetime, the route information is deleted.



Reverse Path (route) Example

A

BC

D E

Source

Destination

Assume same dest_sequence_# for all nodes

A

BC

D E

Source

Destination

Reverse Path FormationReverse Path FormationNetwork LayoutNetwork Layout



Respond to the RREQIf a node has an unexpired entry to the destination in its route tableand

the sequence number associated with that destination is not smaller the that indicated in the RREQ.

To prevent loop and keep freshness.Not respond with older route information than that of source.

Then, the node responds by unicasting a RREP back to the source.If the node cannot satisfy the RREQ, it increments the RREQ’s hop count and then broadcasts the packet to its neighbors.



RREP ProcedureIf the destination node is responding,

Place its current sequence number.Initialize the hop count to zero.Place the length of time this route is valid in the RREP’s lifetime.

If an intermediate node responding,Places its record of the destination’s sequence number.Set the hop count equal to its distance from the destination.Place the amount of time for the destination in the route entry.

source_addr

hop_cnt

dest_addr

source_sequence_#

lifetime (expiration time for reverse path route entry)



RREP ProcedureWhen an intermediate node receives the RREP, it sets up a forward route (path) entry to the destination in its route table.

IP address of the destinationIP address of the neighbor from which the RREP arrivedSequence numberThe hop count to the destinationLifetime contained in the RREP.

So, the forward and reverse route entries of each node can be used for other RREQs.

If a node receives RREPs from more than one neighbor.Forwards the first RREPAnd forwards a later RREP only if that RREP contains a greater sequence number or a smaller hop count.



AODV Route Discovery Example

B

A

S E

F

H

J

D

C

G

IK

Z

Y

Represents a node that has received RREQ for D from S

M

N

L

S D




B

A

S E

F

H

J

D

C

G

IK

Represents transmission of RREQ

Z

YBroadcast transmission

M

N

L




B

A

S E

F

H

J

D

C

G

IK

Represents links on Reverse Path

Z

Y

M

N

L




B

A

S E

F

H

J

D

C

G

IK

• Node C receives RREQ from G and H, but does not forwardit again, because node C has already forwarded RREQ once

Z

Y

M

N

L




B

A

S E

F

H

J

D

C

G

IK

Z

Y

M

N

L




B

A

S E

F

H

J

D

C

G

IK

Z

Y

• Node D does not forward RREQ, because node Dis the intended target of the RREQ

M

N

L




B

A

S E

F

H

J

D

C

G

IK

Z

Y

Represents links on path taken by RREP

M

N

L




B

A

S E

F

H

J

D

C

G

IK

Z

Y

M

N

L

Forward links are setup when RREP travels alongthe reverse path

Represents a link on the forward path




B

A

S E

F

H

J

D

C

G

IK

Z

Y

M

N

L

Routing table entries used to forward data packet.Route is not included in packet header.

DATA



Route MaintenanceIf the source node moves during an active session,

It can reinitiate route discovery to establish a new route to the destination.

When either the destination or some intermediate node moves,A Route Error (RERR) message is sent to the affected source node.When a source node receives the RERR, it can initiate route discovery if the route is still needed.

AB

CD

EF

D’

source

destination

RERRRERR

AB

C

EF

D’

source

destination



AODV Multicast ProtocolUse the same two messages (RREQ and RREP) for multicast route discovery.

Multicast group membership is dynamic: nodes are able to join and leave the group at any time.

Each multicast group has a multicast group leader.Maintain the multicast group sequence numberIt is in no way a central point of failure.core-based tree !!! (group leader = core node)



Multicast Route DiscoveryMulticast route discovery begins either when

A node wishes to join a multicast group (set the Join flag) orWhen it has data to send to a multicast group and does not have a current route to it.

The source node creates RREQwith destination address (multicast group ID)the group’s last known sequence number.Then, broadcast it to it’s neighbors.

The RREQ isA join request, only a node that is a member of the multicast tree may respond.Otherwise, an node with a current route to the multicast group may reply.

Source node means the node that begins route discovery procedure (does not mean multicast group data source)

A node which is not a group member can be a tree member.



Multicast Route DiscoveryIf the RREQ is a join request,

Any node that is not a tree member creates a reverse route entry to the source andBroadcasts the RREQ to it neighbors.A node replies

if it is a router for the multicast group’s tree andIt recorded sequence number for the group is at least as great as that contained in the RREQ.The node unicast RREP back to the source node.

Nodes along the path to the source node that received the RREP,Set up a forward path entry for the group in their multicast route table by adding the node from which they received the RREP as a next hop.Increment the Hop Count field and forward the RREP to the next node.



Multicast Route Discovery

Group Leader

T

T

G

G

G

G

N

G

N

N

N

N

RREQBroadcasting

G: Multicast Group MemberT: Multicast Tree MemberN: Non-Tree Member

N

Group Leader

T

T

G

G

G

G

N

G

N

N

N

N

N

RREQ message propagationRREP sent back to the source



Multicast Route DiscoveryGroup Leader

T

T

G

G

G

G

T

G

N

N

N

N

N

Multicast tree branch addition

Setting the activated flag in its multicast route table.This node is changed into tree member.

These are non-tree members. But now they have route table to the group (i.e., know the best next hop to connect the tree.)

G

Select the greatest multicast group sequence and the smallest hop count to the multicast tree.Source node activates the route by sending a multicast activation (MACT) message.



Leave the treeNon-leaf node

Must not prune itself from the tree, but it must continue to serve as a router for the tree.

Leaf nodeUnicast a MACT message (with the Prune flag set) to its next hop.The next hop

Delete the information for the sending node.If the deletion makes the node a leaf, and the node is not a group member, then it sends a MACT message with the Prune flag set to its next hop.



Leave the tree

Group Leader

T

T

G

G

G

G

T

G

N

N

N

N

N

G MACT (Prune)

MACT (Prune)

Group Leader

T

T

G

G

G

G

G

N

N

N

N

N

GN

GN

Pruning of multicast group member Multicast tree after pruning



Multicast Tree MaintenanceLink breaks

At every hello_interval, a node must receive a broadcast from each of its next hop such as RREQ, Group Hello, data, and Hello (RREP with TTL=1) message.If no broadcast receives during hello_interval from the next hop, then the node send Hello to its next hop.

Failure to receive any broadcasts from the next hop on the multicast tree for hello_life,hello_life=(1+allowed_hello_loss)*hello_intervalIndicates the next hop is out of range and so the link must be repaired.



Multicast Tree MaintenanceWhen a link break occurs

Downstream node is responsible for repairing it.Downstream node broadcasts a join RREQ including an extension field indicating the distance from the group leader.Any node can respond to the RREQ by sending a RREP if it satisfies the following conditions:

It is a part of the multicast treeIt has a fresh enough multicast group sequence number.Its hop count to the leader is smaller than that indicated in the RREQ.

The next procedure is the same as tree join procedure



Multicast Tree Maintenance

Group Leader

T

T

G

G

G

G

G

N

N

N

N

N

GN

GN

Downstreamnode

RREQ (D=2)

RREP

Group Leader

T

T

G

G

G

G

G

N

N

N

N

N

GN

GN

Repaired multicast tree



ZRP Protocol(Zone Routing Protocol)



ZRP BackgroundZRP BackgroundProactive

Nodes store route information before it is needed in routing tables.Little delay to determine route to the destination.Continuously use a large portion of network capacity to keep the routing information current.

ReactiveRouting information is calculated only before data transmissions.Large delay to set up the route.

Hybrid / HierarchicalFor a large networkAssimilate the advantages of both the proactive and reactive protocols.



The Zone Routing Protocol (ZRP)The Zone Routing Protocol (ZRP)A hybrid routing protocol that combines both proactive and on-demand routing strategies

Each node has a predefined zoneInside zones: proactive routing

Intrazone Routing Protocol (IARP)Limits the scope of proactive procedure to node’s local neighborhood

Outside zones: on-demand routingInterzone Routing Protocol (IERP)Search throughout the network is done by a subset of network nodes.

ZRP provides more flexibility



Routing ZonesRouting ZonesRouting Zone:

Defined for each node and includes the nodes whose distance is at most some predefined number.Each node is required to know the topology of the network within its routing zone only and nodes are updated about topology changes only within its routing zone

(using proactive algorithm).

Zone Radius=2S – Central NodeL – outside zoneA-F – NeighborsG-K – Peripheral

The zone is not a description of physical distance but rather of node connectivity (hops).



ZRP ArchitectureZRP Architecture



IntrazoneIntrazone Routing Protocol (IARP)Routing Protocol (IARP)Neighbor Discovery Protocol (NDP)

Used by a node during construction of its routing zone to find its neighboring nodes.

Broadcast “hello” beacons.The quality of the beacon reception is used to indicate the status of the connection to the neighbor.When a neighbor is either lost or found, IARP is notified of this new link status.

Intrazone Routing Protocol (IARP)Nodes broadcast route updates periodically to all nodes in their respective routing zones.Each node learns the topology of its routing zone.

OSPF-like protocols can be used.Reduces update traffic by using the route hop count to prevent updates from propagating beyond the zone radius



InterzoneInterzone Routing Protocol (IERP)Routing Protocol (IERP)Responsible for reactively discovering routes to destinations located beyond a node’s routing zone.

Bordercasting:Packet delivery service that directs messages from one node out to its peripheral nodes.

Make bordercast tree.



IERP Route Discovery OperationIERP Route Discovery Operation1. Source node (S) checks whether the

destination (D) is within its zone. – if yes, send data.

2. S bordercasts a route request to its peripheral nodes (C, H, G).

3. The peripheral nodes execute the same algorithm

4. If the destination is within their zone, a route reply is sent back to the source, indicating the route to the destination.

5. If not, the peripheral nodes forward the route request to their peripheral nodes. (H->B)

B recognizes D is in its routing zone and responds to the route request, indicating the

forwarding path S H B D.

Route accumulation procedure: node appends it address to a received route request packet.

A single route query returns multiple route replies, which can be used to determine the best route based on hop count or any other path metric.



Query Control MechanismsQuery Control MechanismsExcess route request traffic as a result of overlapping queried routing zones

Certain query control mechanisms to be implemented to reduce this amount of query traffic

by steering requests outward from source’s routing zoneand away from each other

Mechanisms Used:Query Detection (QD1/QD2) Early Termination (ET)Selective Bordercasting (SBC)



Query Control Mechanism Query Control Mechanism -- QDQDQuery Detection

When the peripheral nodes continue to bordercast to their peripheral nodes, the query may be relayed through the same nodes again.BRP provides two query detection methods, QD1 and QD2, to notify the remaining nodes through some form of eavesdropping without incurring additional control traffic

QD1Allows intermediate nodes (which relay queries to the edge of the routing zone) to detect queries

QD2In single channel networks, possible for queries to be detected by any node within range of a query transmitting node



QD1/QD2QD1/QD2

Node S bordercasts to two peripheral nodes, B and D

The intermediate nodes A and C can detect the passing packets using QD1.

If QD2 is implemented, node E will be able to receive A’s transmission and record the query as well.



Query Control Mechanism Query Control Mechanism -- Early Early TerminationTermination

Early TerminationA node can prevent a route request from entering already coveredregions through a scheme called early termination.Uses information from query detection to identify which local nodes have already bordercasted the query. A node can prune a peripheral node if it has already relayed a query to that node.



Early TerminationEarly Termination

Node T bordercasts a route request to S and Z through B

B records the request packet (QD1)

S bordercasts to its peripheral node, C

B terminates the transmission to C because it recognizes node C belongs to the previously queried routing zone of node T



Query Control Mechanism Query Control Mechanism -- SBCSBCSelective Bordercasting (SBC)

Rather than bordercast queries to all peripheral nodes, the same coverage can be provided by bordercasting to a chosen subset of peripheral nodes.Requires IARP to provide network topology information for an extended zone that is twice the radius of the routing zone.

A node will first determine the subset of other peripheral nodescovered by its assigned inner peripheral nodes.

The node will then bordercast to this subset of assigned inner peripheral nodes which forms the minimal partitioning set of the outer peripheral nodes.



SBCSBC

S’s inner peripheral nodes are A, B and C

Its outer peripheral nodes are F, G, H, X, Y and Z

2 inner peripheral nodes of B (H and X) are also inner peripheral nodes of A and C

S can choose to eliminate B from its list of bordercastrecipients since A can provide coverage to H and C can cover X



Conclusion RemarksResearch Issues

Multicasting in mobile ad-hoc networksQoS multicasting & Reliable multicastingRouting with QoS constraintsQoS routing with security constrains QoS provisioning network architecturePower-conservative designs in ad-hoc networks

Layer and service dependent power controlPerformance analysis

TCP controlled transferConnection and traffic management

Resource allocation, QoS supports, scheduling, …

9 ad hoc network protocols

Documents

network model

inha universityhttp

inha universityions

krmultim network lab

network overview routing

loss of network connectivity

model of operation ad

infrastructure elements