程序代写代做代考 ER Excel cache algorithm asp 􏸵􏸶

􏸵􏸶
Intro duction
Laptop computers
to show
capacity􏸹 and avail􏹃
continue convenience􏸹 mobility􏸹 memory
improvements
Ad􏹃ho c On􏹃Demand Distance Vector Routing
Charles E􏸶 Perkins
Sun Microsystems Lab oratories Advanced Development Group Menlo Park􏸹 CA 􏸾􏹉􏸻􏸼􏹋
cp erkins􏹀eng􏸶sun􏸶com
Abstract
An ad􏹃ho c network is the cooperative
a col lection of mobile nodes without the
vention of any centralized access point
frastructure􏸶 In this paper we present Ad􏹃hoc On De􏹃
mand Distance Vector Routing
rithm for the operation of such ad􏹃hoc networks􏸶 Each Mobile Host operates as a specialized router􏸹 and routes are obtained as needed 􏹌i􏸶e􏸶􏸹 on􏹃demand􏹍 with little or no reliance on periodic advertisements􏸶 Our new routing algorithm is quite suitable for a dynamic self􏹃 starting network􏸹 as required by users wishing to utilize ad􏹃hoc networks􏸶 AODV provides loop􏹃free routes even while repairing broken links􏸶 Because the protocol does not require global periodic routing advertisements􏸹 the demand on the overal l bandwidth available to the mo􏹃 bile nodes is substantial ly less than in those protocols that do necessitate such advertisements􏸶 Nevertheless we can stil l maintain most of the advantages of basic distance􏹃vector routing mechanisms􏸶 We show that our algorithm scales to large populations of mobile nodes wishing to form ad􏹃hoc networks􏸶 We also include an evaluation methodology and simulation results to verify the operation of our algorithm􏸶
Keywords􏸳 Ad􏹃ho c Networking􏸹 Distance Vector Routing􏸹 Dynamic Routing􏸹 Mobile Networking􏸹 Wire􏹃 less Networks
in
ability of disk storage􏸶 These smaller computers can b e equipp ed with gigabytes of disk storage􏸹 high res􏹃 olution color displays􏸹 p ointing devices􏸹 and wireless communications adapters􏸶 Moreover􏸹 b ecause many of these small 􏹌in size only􏹍 computers op erate with bat􏹃
engagement required inter􏹃 or existing in􏹃
􏹌AODV􏹍􏸹 a novel algo􏹃
Elizab eth M􏸶 Royer
Dept􏸶 of Electrical and Computer Engineering University of California􏸹 Santa Barbara Santa Barbara􏸹 CA 􏸾􏸴􏸵􏸻􏹓 eroyer􏹀alpha􏸶ece􏸶ucsb􏸶edu
tery p ower􏸹 users are free to move ab out at their con􏹃 venience without b eing constrained by wires􏸶
The idea of forming an on􏹃the􏹃􏹘y ad􏹃ho c network of mobile no des dates back to DARPA packet radio net􏹃 work days 􏹆􏸵􏸵􏸹 􏸵􏸼􏹇􏸶 More recently the interest in this sub ject has grown due to availability of license􏹃free􏸹 wireless communication devices that users of laptop computers can use to communicate with each other􏸶 Several recent pap ers on this topic have fo cused on the algorithmic complexity of cho osing the optimal set of ad􏹃ho c routers 􏹆􏹓􏸹 􏸺􏸹 􏸵􏹋􏹇􏸹 while others have prop osed new routing solutions 􏹆􏹉􏸹 􏸽􏸹 􏸵􏸻􏸹 􏸵􏹉􏸹 􏸵􏹓􏸹 􏸵􏸺􏹇 leverag􏹃 ing features from the existing Internet routing algo􏹃 rithms􏸶 Interest within the Internet Engineering Task Force 􏹌IETF􏹍 is also growing as is evidenced by the for􏹃 mation of a new working group 􏹌manet 􏹆􏹋􏸹 􏸵􏸴􏹇􏹍 whose charter is to develop a solution framework for rout􏹃 ing in ad􏹃ho c networks􏸶 The manet working group has goals that are quite distinct from the goals of the IETF mobileip working group􏸹 and make little or no use of Mobile IP 􏹆􏸼􏸻􏹇 or any of its forerunners 􏹌e􏸶g􏸶􏸹 􏹆􏸾􏸹 􏸼􏸼􏹇􏹍􏸶
The Destination􏹃Sequenced Distance Vector 􏹌DSDV􏹍 algorithm has b een prop osed 􏹆􏸵􏸺􏹇 as a variant of the distance vector routing metho d by which mobile no des co op erate to form an ad􏹃ho c network􏸶 DSDV is e􏹂ective for creating ad􏹃ho c networks for small p opu􏹃 lations of mobile nodes􏸹 but it is a fairly brute force approach because it depends for its correct operation on the periodic advertisement and global dissemina􏹃 tion of connectivity information􏸶 Frequent system􏹃wide broadcasts limit the size of ad􏹃ho c networks that can e􏹂ectively use DSDV b ecause the control message over􏹃 head grows as O 􏹌n􏸼 􏹍􏸶 DSDV also requires each mobile no de to maintain a complete list of routes􏸹 one for each destination within the ad􏹃ho c network􏸶 This almost al􏹃 ways exceeds the needs of any particular mobile no de􏸶 Keeping a complete routing table do es reduce route ac􏹃 quisition latency b efore transmission of the 􏸷rst packet to a destination􏸶 It is􏸹 however􏸹 p ossible to design a sys􏹃
of

tem whereby routes are created on􏹃demand 􏹌e􏸶g􏸶􏸹 􏹆􏸵􏸻􏹇􏹍􏸶 Such systems must take steps to limit the time used for route acquisition􏹈 otherwise􏸹 users of the ad􏹃ho c no des might exp erience unacceptably long waits b efore trans􏹃 mitting urgent information􏸶 The advantage here is that a smo othly functioning ad􏹃ho c system with on􏹃demand routes could largely eliminate the need for p erio dic broadcast of route advertisements􏸶 With the goals of minimizing broadcasts and transmission latency when new routes are needed􏸹 we designed a proto col to im􏹃 prove up on the p erformance characteristics of DSDV in the creation and maintenance of ad􏹃ho c networks􏸶
Although AODV do es not dep end sp eci􏸷cally on particular asp ects of the physical medium across which
nicate􏸹 unless the former node is o􏹂ering its services as an intermediate forwarding station to maintain con􏹃 nectivity b etween two other no des􏸶
When the lo cal connectivity of the mobile no de is of interest􏸹 each mobile no de can b ecome aware of the other nodes in its neighborhood by the use of several techniques􏸹 including local 􏹌not system􏹃wide􏹍 broad􏹃 casts known as hello messages􏸶 The routing tables of the no des within the neighb orho o d are organized to op􏹃 timize resp onse time to lo cal movements and provide quick resp onse time for requests for establishment of new routes􏸶 The algorithm􏹐s primary objectives are􏸳
􏹚 To broadcast discovery packets only when neces􏹃 sary
􏹚 To distinguish b etween lo cal connectivity manage􏹃
packets
largely
such as
wireless
a mobile no de can have neighb ors which hear its broad􏹃 casts and yet do not detect each other 􏹌the hidden ter􏹃 minal problem 􏹆􏸼􏸵􏹇􏹍􏸶 We do not make any attempt to
are disseminated􏸹 its development has b een motivated by limited range broadcast media those utilized by infrared or radio frequency communications adapters􏸶 Using such media􏸹
ment 􏹌neighborhood detection􏹍 and
ogy maintenance
􏹚 To disseminate information ab out
general top ol􏹃
use
our
by
ters􏸶 No des that need to op erate over multiple chan􏹃 nels are presumed to be able to do so􏸶 The algorithm works on wired media as well as wireless media􏸹 as long as links along which packets may b e transmitted are
that are likely to need the information􏸶
AODV uses a broadcast route discovery mecha􏹃 nism 􏹆􏹉􏹇􏸹 as is also used 􏹌with mo di􏸷cations􏹍 in the Dy􏹃 namic Source Routing 􏹌DSR􏹍 algorithm 􏹆􏸵􏸻􏹇􏸶 Instead of source routing􏸹 however􏸹 AODV relies on dynam􏹃 ically establishing route table entries at intermediate no des􏸶 This di􏹂erence pays o􏹂 in networks with many no des􏸹 where a larger overhead is incurred by carry􏹃 ing source routes in each data packet􏸶 To maintain the most recent routing information b etween no des􏸹 we b orrow the concept of destination sequence num􏹃 b ers from DSDV 􏹆􏸵􏸺􏹇􏸶 Unlike in DSDV􏸹 however􏸹 each ad􏹃ho c no de maintains a monotonical ly increasing se􏹃 quence numb er counter which is used to sup ersede stale cached routes􏸶 The combination of these techniques yields an algorithm that uses bandwidth e􏹊ciently 􏹌by minimizing the network load for control and data traf􏹃 􏸷c􏹍􏸹 is resp onsive to changes in top ology􏸹 and ensures lo op􏹃free routing􏸶
2.1. Path Discovery
sp eci􏸷c characteristics of the physical medium in
algorithm􏸹 nor to handle sp eci􏸷c problems p osed channelization needs of radio frequency transmit􏹃
available􏸶 The only requirement cast medium is that neighb oring others􏹐 broadcasts􏸶
placed on no des can
the broad􏹃 detect each
AODV uses
no des􏸶 It do es
no des when one of the no des cannot hear the other one􏹈 however we may include the use of such links in future enhancements􏸶 Steps to prevent use of such asymmet􏹃 ric links b etween no des are describ ed brie􏹘y in Sec􏹃 tion 􏸼􏸶􏹉􏸶
symmetric links not attempt to
b etween
follow paths b etween
The remainder of this
In Section 􏸼􏸹 the proto col details
Section 􏸴 presents the simulations􏸹 input parameters􏸹 and results obtained􏸶 Section 􏹉 describ es our plans for future work􏸹 and 􏸷nally Section 􏹋 concludes the pap er􏸶
􏸼􏸶 The Ad􏹃ho c On􏹃Demand Vector Algorithm
Our basic prop osal can b e
Distance
pap er is organized as follows􏸶 for AODV are given􏸶
route acquisition system􏹈
tive paths neither maintain any routing information nor participate in any p erio dic routing table exchanges􏸶 Further􏸹 a no de do es not have to discover and maintain a route to another no de until the two need to commu􏹃
the
following 􏸷elds􏸳
neighb oring
on􏹃demand
􏹌RREQ􏹍 packet to its
neighb ors􏸶
The RREQ
contains
changes in lo􏹃 cal connectivity to those neighb oring mobile no des
The Path
source no de
for which it has no routing information in its table􏸶
Discovery pro cess is initiated whenever a needs to communicate with another no de
Every no de
sequence number and a broadcast id􏸶 The source no de initiates path discovery by broadcasting a route request
maintains two separate counters􏸳 a node
called a pure
no des that do not lie on ac􏹃
sour ce seq uence 􏹑􏹈 br oadcast id􏹈 dest addr􏹈 dest seq uence 􏹑􏹈 hop cnt 􏹛
􏺇 sour ce addr􏹈
􏸼

The pair 􏺇 sour ce addr􏹈 br oadcast id 􏹛 uniquely identi􏸷es a RREQ􏸶 br oadcast id is incremented when􏹃 ever the source issues a new RREQ􏸶 Each neighb or either satis􏸷es the RREQ by sending a route reply 􏹌RREP􏹍 back to the source 􏹌see Section 􏸼􏸶􏸵􏸶􏸼􏹍􏸹 or re􏹃 broadcasts the RREQ to its own neighb ors after in􏹃 creasing the hop cnt􏸶 Notice that a no de may receive multiple copies of the same route broadcast packet from
As the RREQ travels from a source to various desti􏹃 nations􏸹 it automatically sets up the reverse path from
all no des back to the source
ure 􏸵􏸶 To set up a reverse
address of the neighb or from which it received the 􏸷rst copy of the RREQ􏸶 These reverse path route entries are maintained for at least enough time for the RREQ to traverse the network and pro duce a reply to the sender􏸶
2.1.2. Forward Path Setup
􏹆􏹉􏹇􏸹 as illustrated in Fig􏹃 path􏸹 a no de records the
various neighb ors􏸶
a RREQ􏸹 if it has
same br oadcast id and source address􏸹 it drops the re􏹃 dundant RREQ and do es not rebroadcast it􏸶 If a no de cannot satisfy the RREQ􏸹 it keeps track of the follow􏹃
When an intermediate no de receives already received a RREQ with the
ing information in order to implement the setup􏸹 as well as the forward path setup
reverse path
company the transmission of the
eventual
􏹚 Destination IP address
􏹚 Source IP address
􏹚 Broadcast id
􏹚 Expiration time for reverse path route entry
that will RREP􏸳
ac􏹃
RREQ was received over a bi􏹃directional intermediate no de has a route entry for the
link􏸶 If desired
Eventually􏸹 a RREQ will arrive at a
the destination itself 􏹍 that p ossesses a current route to the destination􏸶 The receiving no de 􏸷rst checks that
the
an
destination􏸹 it determines whether the route is current by comparing the destination sequence numb er in its own route entry to the destination sequence numb er in the RREQ􏸶 If the RREQ􏹐s sequence numb er for the destination is greater than that recorded by the inter􏹃 mediate no de􏸹 the intermediate no de must not use its recorded route to resp ond to the RREQ􏸶 Instead􏸹 the
no de
􏹌p ossibly
􏹚 Source
no de􏹐s sequence
D
numb er􏸶
D
intermediate no de
mediate node can reply only when it has a route with a sequence numb er that is greater than or equal to that contained in the RREQ􏸶 If it do es have a current route to the destination􏸹 and if the RREQ has not b een pro cessed previously􏸹 the no de then unicasts a route re􏹃
rebroadcasts the RREQ􏸶 The inter􏹃
ply packet 􏹌RREP􏹍 back to its received the RREQ􏸶 A RREP information􏸳
neighb or from which it contains the following
􏺇 sour ce addr􏹈 dest addr􏹈 dest seq uence 􏹑􏹈 hop cnt􏹈 l if etime 􏹛
Figure 1. Reverse Path Formation
2.1.1. Reverse Path Setup
Figure 2. Forward Path Formation
There are two sequence
the broadcast id􏹍 included in a RREQ􏸳 the source se􏹃 quence numb er and the last destination sequence num􏹃 b er known to the source􏸶 The source sequence numb er is used to maintain freshness information about the re􏹃 verse route to the source􏸹 and the destination sequence numb er sp eci􏸷es how fresh a route to the destination must b e b efore it can b e accepted by the source􏸶
numb ers 􏹌in
addition
to
arrives at a no de destination􏸹 a re􏹃 to the source of the RREP travels back no de along the path sets up a the no de from which the RREP timeout information for route en􏹃 tries to the source and destination􏸹 and records the latest destination sequence numb er for the requested destination􏸶 Figure 􏸼 represents the forward path setup as the RREP travels from the destination D to the source no de S􏸶 No des that are not along the path determined by the RREP will timeout after AC􏹃 TIVE ROUTE TIMEOUT 􏹌􏸴􏸻􏸻􏸻 msec􏹍 and will delete
p ointers􏸶
receiving an RREP propagates the 􏸷rst
a given source no de towards that source􏸶 If it receives further RREPs􏸹 it up dates its routing in􏹃 formation and propagates the RREP only if the RREP
S
timeout
By the time a broadcast packet that can supply a route to the
S
verse path has b een established RREQ 􏹌Section 􏸼􏸶􏸵􏸶􏸵􏹍􏸶 As the to the source􏸹 each
forward p ointer to
came􏸹 up dates its
􏸴
the reverse A no de RREP for

contains either a greater destination sequence numb er than the previous RREP􏸹 or the same destination se􏹃 quence numb er with a smaller hop count􏸶 It suppresses all other RREPs it receives􏸶 This decreases the num􏹃 b er of RREPs propagating towards the source while also ensuring the most up􏹃to􏹃date and quickest routing information􏸶 The source no de can b egin data trans􏹃 mission as soon as the 􏸷rst RREP is received􏸹 and can later update its routing information if it learns of a b etter route􏸶
2.2. Route Table Management
Each time a route entry is used to transmit data from a source toward a destination􏸹 the timeout
which the route is considered to b e invalid􏸶 routing table entry􏸹 the address of active through which packets for the given desti􏹃 received is also maintained􏸶 A neighb or is considered active 􏹌for that destination􏹍 if it origi􏹃 nates or relays at least one packet for that destination within the most recent activ e timeout p erio d􏸶 This in􏹃 formation is maintained so that all active source no des can be noti􏸷ed when a link along a path to the des􏹃 tination breaks􏸶 A route entry is considered active if it is in use by any active neighb ors􏸶 The path from a source to a destination􏸹 which is followed by pack􏹃 ets along active route entries􏸹 is called an active path􏸶 Note that􏸹 as with DSDV􏸹 all routes in the route table are tagged with destination sequence numb ers􏸹 which guarantee that no routing lo ops can form􏸹 even under extreme conditions of out􏹃of􏹃order packet delivery and
􏹌metric􏹍
destination route
􏹉
for the entry is reset to the current time
plus ac􏹃
the mo􏹃 numb er numb er
for the current route􏸶 The route with the greater se􏹃 quence numb er is chosen􏸶 If the sequence numb ers are the same􏸹 then the new route is selected only if it has
tive If
bile
route timeout􏸶
a new route is no de compares
In addition to the source and
numb ers􏸹 other useful information is also stored in the route table entries􏸹 and is called the soft􏹃state asso􏹃 ciated with the entry􏸶 Asso ciated with reverse path routing entries is a timer􏸹 called the route request ex􏹃 piration timer􏸶 The purp ose of this timer is to purge reverse path routing entries from those no des that do not lie on the path from the source to the destination􏸶 The expiration time dep ends up on the size of the ad􏹃 ho c network􏸶 Another imp ortant parameter asso ciated with routing entries is the route caching timeout􏸹 or the
time after In each
neighb ors nation are
destination
sequence
o􏹂ered to a mobile no de􏸹 the destination sequence of the new route to the destination sequence
a smaller metric 􏹌fewer numb er of nation􏸶
2.3. Path Maintenance
hops􏹍 to the
desti􏹃
Movement of no des not lying
does not a􏹂ect the routing to that path􏹐s destination􏸶 If the source no de moves during an active session􏸹 it can reinitiate the route discovery pro cedure to estab􏹃 lish a new route to the destination􏸶 When either the destination or some intermediate no de moves􏸹 a sp ecial RREP is sent to the a􏹂ected source nodes􏸶 Periodic hel lo messages can b e used to ensure symmetric links􏸹 as well as to detect link failures􏸹 as describ ed in Sec􏹃 tion 􏸼􏸶􏹉􏸶 Alternatively􏸹 and with far less latency􏸹 such failures could b e detected by using link􏹃layer acknowl􏹃 edgments 􏹌LLACKS􏹍􏸶 A link failure is also indicated if attempts to forward a packet to the next hop fail􏸶
Once the next hop b ecomes unreachable􏸹 the no de upstream of the break propagates an unsolicited RREP with a fresh sequence numb er 􏹌i􏸶e􏸶􏸹 a sequence numb er that is one greater than the previously known sequence numb er􏹍 and hop count of 􏸵 to all active upstream neighb ors􏸶 Those no des subsequently relay that mes􏹃 sage to their active neighb ors and so on􏸶 This pro􏹃 cess continues until all active source no des are noti􏸷ed􏹈 it terminates b ecause AODV maintains only lo op􏹃free routes and there are only a 􏸷nite numb er of no des in the ad􏹃ho c network􏸶
Up on receiving noti􏸷cation of a broken link􏸹 source no des can restart the discovery pro cess if they still re􏹃 quire a route to the destination􏸶 To determine whether
route
level
tions
the source no de 􏹌or any other no de along the previ􏹃 ous route􏹍 decides it would like to rebuild the route to the destination􏸹 it sends out an RREQ with a desti􏹃 nation sequence numb er of one greater than the previ􏹃 ously known sequence numb er􏸹 to ensure that it builds a new􏸹 viable route􏸹 and that no nodes reply if they still regard the previous route as valid􏸶
has b een
proto col
remain op en using the
used recently􏸹 control blo cks
as well as insp ect upp er􏹃 to see whether connec􏹃 indicated destination􏸶 If
along an
active path
high no de mobility 􏹌see App endix A􏹍􏸶
A mobile no de maintains a route table entry for each
destination of interest􏸶 Each route table entry contains the following information􏸳
􏹚 Destination
􏹚 Next Hop
􏹚 Numb er of hops
􏹚 Sequence numb er for the
􏹚 Active neighb ors for this
􏹚 Expiration time for the route table entry
a route is still
needed􏸹 a no de may check whether the

S DATA
VOICE
Simulated proto col
Packet size 􏹌bytes􏹍
Packet count
Inter􏹃arrival time of data packets Session interval 􏹌sec􏹍
UDP
􏹓􏹉
Exp onential􏹃mean 􏸵􏸻􏸻􏸻 􏸼􏸻 msec Geometric􏹃mean 􏸾􏸻􏸻
UDP
􏸵􏸽􏸻
Exp onential􏹃mean 􏸵􏸻􏸻􏸻 􏸼􏸻 msec Geometric􏹃mean 􏹓􏸻􏸻
2.4. Local Connectivity Management
No des learn of their neighb ors in one of two ways􏸶
develop ed at UCLA as the successor to Maisie􏹆􏸼􏹇􏸶 The PARSEC language is suited to the simulation of dy􏹃 namic top ologies and routing algorithms􏸶
The main ob jective of our simulations is to show that on􏹃demand route establishment with AODV is both quick and accurate􏸶 Additional objectives include showing that AODV scales well to large networks􏸹 and determining the optimal value for each of the necessary parameters􏸶
3.1. Simulation Environment
Our simulations were run using networks of 􏹋􏸻􏸹 􏸵􏸻􏸻􏸹 􏹋􏸻􏸻􏸹 and 􏸵􏸻􏸻􏸻 no des􏸶 The movement algorithm for all network sizes is the same􏸶 No des are initially placed randomly within a 􏸷xed􏹃size L 􏹠 L area􏸶 During the simulation􏸹 no des are free to move anywhere within this area􏸶 Each no de cho oses a sp eed from a uniform distri􏹃 bution b etween 􏸻􏸶􏹉 and 􏸻􏸶􏸺 meters p er second􏸶 It then travels towards a random sp ot within the L 􏹠 L area􏸶 The no de moves until it reaches that sp ot􏸹 then cho oses a rest p erio d from a uniform distribution b etween 􏹓􏸻
from a neigh􏹃 information to In the event all of its ac􏹃 interval􏸹 it broadcasts to its neighb ors a hel lo message 􏹌a sp e􏹃 cial unsolicited RREP􏹍􏸹 containing its identity and sequence numb er􏸶 The no de􏹐s sequence numb er is not changed for hel lo message transmissions􏸶 This
Whenever a no de receives a broadcast b or􏸹 it up dates its lo cal connectivity
ensure
that a
tive downstream neighb ors within hello
that it includes this neighb or􏸶 no de has not sent any packets to
sages are not received
tive path􏸹 the active
are sent noti􏸷cation
Section 􏸼􏸶􏸴􏸶 We have determined the for allowed hello loss is two􏸹 as is tion 􏸴􏸶􏸼􏸶
seconds􏸶 After the rest p erio d􏸹 the no de travels another randomly selected sp ot􏸶 This pro cess throughout the simulation􏸹 causing continuous
Table 1. Session-Dependent Traffic Parameters.
hel lo
side
tains
that
ity information
or a hel lo from
allowed hello loss consecutive hel lo messages from a no de previously in the neighb orho o d􏸹 is an indica􏹃 tion that the lo cal connectivity has changed􏸶 Fail􏹃 ing to receive hel lo messages from inactive neighbors do es not trigger any proto col action􏸶 If hel lo mes􏹃
message is prevented from being rebroadcast out􏹃 the neighb orho o d of the no de b ecause it con􏹃
a time to live 􏹌TTL􏹍 value of 􏸵􏸶 receive this packet up date their lo cal to the no de􏸶 Receiving a
Neighb ors connectiv􏹃 broadcast a new neighbor􏸹 or failing to receive
from the next hop along an ac􏹃 neighb ors using that next hop of link failure as describ ed in
optimal value shown in Sec􏹃
The local connectivity management with hel lo mes􏹃
also uses the same channel transmission􏸹 carrier sens􏹃 ing is p erformed by a no de to determine whether any of its neighb ors is transmitting􏸶 If the no de detects an ongoing transmission by a neighb or􏸹 it calculates an ex􏹃 p onential backo􏹂 based on the numb er of times it has attempted the retransmission and waits this amount of
and 􏸴􏸻􏸻
towards
rep eats
changes in the top ology of the underlying network􏸶
Each of the simulations
mo del􏸶
time b efore listening to the tempts to transmit a packet dropping the packet􏸶
Before
b eginning a
sages can also b e used to
bidirectional connectivity are considered to b e neigh􏹃 b ors􏸶 For this purp ose􏸹 each hel lo sent by a no de lists the no des from which it has heard􏸶 Each no de checks to make sure that it uses only routes to neighb ors that have heard the node􏹐s hel lo message􏸶 To save local
ensure that only no des with
We have simulated AODV using an event􏹃driven􏸹 packet􏹃level simulator called PARSEC􏹆􏸵􏹇􏸹 which was
􏹋
simultaneously
packets collide at no de B and b oth packets are dropp ed􏸶
Each no de creates a session to another no de selected at random􏸶 The sessions created for each simulation
channel again􏸶 A no de at􏹃 max retrans times b efore
bandwidth􏸹 such checking should b e p erformed explicitly con􏸷gured into the no des􏸶
􏸴􏸶 Simulations and Results
only if
No des in the
hidden terminal
B􏸹 and no de C􏸹 unable to hear no de A􏹐s transmission􏸹
simulation
problem􏸶 If no de A transmits to no de
frequently su􏹂er from the
transmits to no de B􏸹 we assume the

are of
􏹌S DATA􏹍 packets or voice
each of the session typ es are given in Table 􏸵􏸶 We chose to use the small data packet sessions for most of our simulations b ecause the larger size of the voice packets and the greater numb er of sessions generated tended to congest the network and hence decrease the go o dput ratio􏸶 Nevertheless􏸹 we include the results from these simulations to o􏹂er a contrast to the lighter demands
homogeneous
of the small data packets and to place a greater on the proto col􏸶
typ e􏹈
they are
data􏸶 The parameters for
A session sends data segments until
sent the desired numb er of segments or it receives a
either
small data
either
it has
stress
we did not
The interconnection pattern of an ad􏹃ho c network is determined in part by the communication range 􏹌Rmax 􏹍􏸶 For our simulations􏸹 we held Rmax constant at 􏸵􏸻m􏸶 Two no des can communicate directly􏸹 and are
Hello Interval
Route Discovery Timeout
Route Expiration
Reverse Route Life
Maximum 􏹑 of Retransmissions
􏸵􏸻􏸻􏸻 msec
􏸵􏸻􏸻􏸻 msec
􏸴􏸻􏸻􏸻 msec
􏸴􏸻􏸻􏸻 msec 􏸵􏸻
thus considered each other􏹐s neighb ors􏸹
than Rmax distance apart􏸶 The ro om
and 􏸵􏸻􏸻 no des networks is 􏹋􏸻m􏹠􏹋􏸻m􏸶
we found 􏹋􏸻m􏹠􏹋􏸻m to b e to o small􏸹 so we increased the dimensions to 􏸵􏸻􏸻m􏹠􏸵􏸻􏸻m􏸶 Similarly􏸹 for 􏸵􏸻􏸻􏸻 no des we used a ro om size of 􏸵􏹋􏸻m􏹠􏸵􏹋􏸻m􏸶
Table 􏸼 gives the values of the essential parameters for AODV􏸶 The parameter values were chosen b ecause
3.2. Results and Discussion
and allowed hello loss􏸹 we varied
ob jective was to show quickly and accurately􏸶 at this time know an optimal value
that AODV can
Our 􏸷rst 􏸷nd routes
Since
for rreq retries rreq retries b e􏹃
if they are less size for the 􏹋􏸻 For 􏹋􏸻􏸻 no des􏸹
Table 2. Simulated Parameter Values
they minimize network congestion while
allowing the accurately as
algorithm to p ossible􏸶
op erate
as quickly
and as
a we intuitively guessed would b e reasonable􏸶 Fig􏹃 ure 􏸴 shows the go o dput ratios for 􏹋􏸻 and 􏸵􏸻􏸻 no des
tween value
􏸻 and 􏸴 and set allowed hello loss to 􏸼􏸹
timeout message from the network layer􏸶
triggered when a no de has sent a RREQ for a particular destination and has not received a valid route within
available during a session􏸹 packets are network layer􏸶 The data rate for b oth 􏸵􏸶􏸻 Mbit􏸿sec􏸶
Timeouts are
using
go o dput ratio is consistently ab ove 􏸾􏸺􏹖􏸶 For 􏸵􏸻􏸻 no des􏸹 the go o dput ratio for rreq retries􏹎􏸻 is ap􏹃 proximately 􏸾􏹓􏹖􏸹 but then it decreases to 􏸾􏸼􏹖 for rreq retries􏹎􏸵 and then increases with increasing values of rreq retries􏸶 Bro ch et al􏸶 􏹆􏸴􏹇 simulated AODV over a network of 􏹋􏸻 no des and achieved go o d􏹃 put ratios b etween 􏸾􏸽􏹖 and 􏸵􏸻􏸻􏹖􏸹 dep ending on the amount of time the no des were stationary during the simulation􏸶 Note that our S DATA simulation uses the same size data packets as they did􏸶 Hence􏸹 our
the S DATA session typ e􏸶 For 􏹋􏸻 no des􏸹 the
route discovery timeout􏸶 Any time a route
is not by the session typ es is
dropp ed
Each simulation is run for 􏹓􏸻􏸻 seconds􏸹 and new ses􏹃 sions are generated throughout the simulation􏸶 Hence􏸹
we keep track of􏸹 and account for􏸹
any uncompleted at the end of the
sessions and data simulation􏸶
packets in transit
100
99
98
97
96
95
94
93
92
91
90
50 Nodes 100 Nodes
0 0.5 1 1.5 2 2.5 3
RREQ_RETRIES
100
99
98
97
96
95
94
93
92
91
90
50 Nodes 100 Nodes
0 0.5 1 1.5 2 2.5 3 ALLOWED_HELLO_LOSS
Figure 3. Achieved Goodput for Varying
rreq retries
Figure 4. Achieved Goodput for Varying
allowed hello loss
􏹓
Goodput(%)
Goodput(%)

300
250
200
150
100
50
0
50 Nodes 100 Nodes
0 0.5 1 1.5 2 2.5 3 RREQ_RETRIES
600
500
400
300
200
100
0
50 Nodes 100 Nodes
0 0.5 1 1.5 2 2.5 3 ALLOWED_HELLO_LOSS
Figure 5. Route Acquisition Latency for Varying rreq retries
achieved go o dput ratio for a 􏹋􏸻 no de network roughly corresponds with their results for the same size net􏹃 work􏸹 with our results b eing slightly b etter􏸶 We disre􏹃 gard the arti􏸷cially high go o dput ratio for 􏸵􏸻􏸻 no des and rreq retries􏹎􏸻 b ecause 􏸼􏸻􏹖 more of the ses􏹃 sions ab orted in this simulation than in the simulations with larger rreq retries values􏸶 Given the remaining go o dput ratios for 􏹋􏸻 and 􏸵􏸻􏸻 no des􏸹 we set the optimal rreq retries value to 􏸼􏸶
Figure 6. Route Acquisition Latency for Varying allowed hello loss
􏸷rst RREQ was sent􏸶 If a route to a destination was never found􏸹 this time lapse was not taken into account in the computation􏸶 Figure 􏹋 shows the computed route acquisition latencies for varying rreq retries values􏸹 and Figure 􏹓 shows the corresp onding values for varying allowed hello loss values􏸶 With the exception of rreq retries 􏹎 􏸻􏸹 the minimum route acquisition latency was attained for the combination rreq retries􏹎􏸼􏸹 allowed hello loss􏹎􏸼􏸹 giving fur􏹃 ther credence to our choice of parameter values􏸶
We with
hello
then simulated rreq retries􏹎􏸼 loss parameter􏸶
􏹋􏸻 and 􏸵􏸻􏸻 no des networks
the allowed of these simu􏹃 lations are shown in Figure 􏹉􏸶 Here􏸹 for 􏹋􏸻 no des􏸹 allowed hello loss􏹎􏸼 pro duced the b est results􏸹 while for 􏸵􏸻􏸻 no des allowed hello loss􏹎􏸻 was the b est􏸶 Again􏸹 b ecause 􏸻 is an unrealistic value􏸹 and b ecause allowed hello loss􏹎􏸼 pro duced the second
􏸼 to b e
the essential results of our simulations
and varied The results
Table 􏸴 gives
for networks of 􏹋􏸻􏸹 􏸵􏸻􏸻􏸹 􏹋􏸻􏸻􏸹 and 􏸵􏸻􏸻􏸻 no des􏸶 The re􏹃 sults were obtained using the S DATA session typ e􏸹 and setting rreq retries􏹎􏸼 and allowed hello loss􏹎􏸼􏸶 The bandwidth overhead ratio is a metric taken from 􏹆􏸵􏸽􏹇 􏹌although there it is called bandwidth utilization􏹍􏸹 and is computed by dividing the total numb er of bits
b est results􏸹
the optimal
􏸷nding that
ter p erformance􏸶 In their simulations
numb er of data bits trans􏹃
we chose
value􏸶 This contradicts
allowed hello loss 􏹎 Bro ch
calculation
b ecause it
gives over􏹃 b oth
transmitted by the total
mitted􏸶 We include this
a go o d representation of the amount of control
head asso ciated with a proto col􏸶 We also rep ort
the instantaneous go o dput ratio at the simulation end􏸹 as well as the average go o dput ratio throughout the simulation􏸹 b ecause these numb ers can vary due to a sudden increase in link breakages or session cre􏹃 ations at the end of the simulation􏸶 The 􏸵􏸻􏸻􏸻 no de simulation was run for a shorter p erio d of time b e􏹃 cause of the di􏹊culty of running such a large simu􏹃 lation􏸶 Also􏸹 the 􏹋􏸻􏸻 and 􏸵􏸻􏸻􏸻 no de simulations had a slightly larger session generation interval than the 􏹋􏸻 and 􏸵􏸻􏸻 no des networks in order to keep the to􏹃
et al􏸶􏹐s pro duces b et􏹃 they also used of the two pa􏹃 rameters may account for the slightly decreased go o d􏹃 put that their AODV simulations pro duced􏸶 Another signi􏸷cant di􏹂erence b etween their simulation and ours is that they set ROUTE DISCOVERY TIMEOUT to 􏹓􏸻􏸻􏸻 msec􏸹 whereas we found the optimal value to b e
hello loss 􏹎 􏸴 􏹎 􏸴􏸶 The combination
allowed
rreq retries
􏸵􏸻􏸻􏸻 msec􏸶
To show that AODV 􏸷nds
routes in a timely man􏹃 ner􏸹 we examined the route acquisition latency􏸶 The route acquisition latency was computed by noting the simulation time when an initial RREQ was broadcast for a given destination􏸹 and then noting the time when the 􏸷rst RREP was received at the source􏸶 For suc􏹃 cessive RREQ retries for the same route􏸹 the start time for the route was held at the time at which the
We note networks Reasons decreased go o dput ratio are a much greater collision rate􏸹 due to the increase in the numb er of
􏸽
tal numb er of sessions more manageable􏸶
that the results of the
􏹋􏸻􏸻 and 􏸵􏸻􏸻􏸻 no de would have desired􏸶
are not for the
as go o d as we
Route Acquisition Latency (msec)
Route Acquisition Latency (msec)

􏹑 of No des
􏹋􏸻
􏸵􏸻􏸻
􏹋􏸻􏸻
􏸵􏸻􏸻􏸻
Go o dput Ratio at sim end
􏸾􏸺􏸶􏸽􏹋􏹖
􏸾􏸴􏸶􏸾􏸼􏹖
􏸺􏸽􏸶􏹉􏹓􏹖
􏸽􏸻􏸶􏹋􏸴􏹖
Go o dput Ratio avg
􏸾􏸽􏸶􏸾􏸺􏹖
􏸾􏹋􏸶􏸾􏸵􏹖
􏸺􏹓􏸶􏹉􏸴􏹖
􏸽􏸼􏸶􏸴􏸼􏹖
Bandwidth Overhead Ratio
􏸵􏸶􏸵􏹉
􏸵􏸶􏸵􏸵
􏸵􏸶􏸴􏸵
􏸵􏸶􏹉􏸾
Avg Rte Acq Latency 􏹌msec􏹍
􏸼􏸻􏹓
􏸼􏸻􏸼
􏹉􏹋􏹉
􏹋􏹉􏸺
Avg Path Length 􏹌hops􏹍
􏸴􏸶􏸾􏹉
􏹉􏸶􏹋􏸽
􏹓􏸶􏸺􏸴
􏸵􏸻􏸶􏹉􏹋
Loss to Collision
􏸵􏸶􏹉􏸴􏹖
􏹋􏸶􏸽􏹉􏹖
􏸼􏸼􏸶􏸺􏸻􏹖
􏸼􏹓􏸶􏸴􏸽􏹖
Ro om Size 􏹌m􏹍
􏹋􏸻x􏹋􏸻
􏹋􏸻x􏹋􏸻
􏸵􏸻􏸻x􏸵􏸻􏸻
􏸵􏹋􏸻x􏸵􏹋􏸻
Simulation Length 􏹌sec􏹍
􏹓􏸻􏸻
􏹓􏸻􏸻
􏹓􏸻􏸻
􏸴􏸻􏸻
􏹑 Generated Sessions
􏸼􏹉
􏹓􏸼
􏸵􏸽􏸼
􏸼􏹓􏸴
􏹑 Completed Sessions
􏸼􏸵
􏹉􏹓
􏸵􏸵􏸽
􏸵􏸼􏸻
􏹑 Ab orted Sessions
􏸻
􏸼
􏸴􏸼
􏸺􏸴
Table 3. Summary of S DATA Results
nodes and the longer paths causing a greater like􏹃 lihood for collisions during the hop􏹃by􏹃hop forward􏹃 ing of the message􏸹 and the added interference of all the hello messages􏸶 Also􏸹 the route acquisition la􏹃 tency increased due to the larger average path length
and the
mission
access􏸶
mance
scalable ad􏹃ho c routing
networks this large we are pushing the
bilities of mobile networks􏸹 as we are not aware of any other attempts to mo del networks of such a large size􏸶
We also ran simulations of the 􏹋􏸻 and 􏸵􏸻􏸻 no de net􏹃
of
same amount of control overhead 􏹌i􏸶e􏸶 the same numb er of RREQs and RREPs􏹍􏸹 the voice sessions send many more data bits b ecause of the increased data packet size􏸶
additional
b ecause of increased However􏸹 regardless
delay in control message trans􏹃 comp etition for channel
􏹉􏸶
net Draft 􏹆􏸵􏸾􏹇 submitted
group􏸶 There are a numb er of
which may supp ort larger p opulations of ad􏹃ho c users􏸹
of the values􏸹 AODV is currently
decreased p erfor􏹃 one of the most
proto cols􏸶 We
feel that with current capa􏹃
Inter􏹃
typ e describ ed in Sec􏹃 typ e to stress the abil􏹃 these simulations􏸹 to􏹃 results from the S DATA Table 􏹉􏸶 The two im􏹃 portant results from these simulations are the good􏹃 put ratio and the bandwidth overhead ratio􏸶 The go o dput ratio for the voice session typ e was lower than that of the S DATA sessions􏸶 This is due to the fact that there were signi􏸷cantly more collisions due to the longer data packet lengths􏸶 Also􏸹 b e􏹃 cause the data packets were larger and to ok longer to transmit􏸹 we found that the queues of the no des frequently backed up b ecause they had to wait for channel access for p ossibly lengthy p erio ds of time􏸹 causing delays in sending RREQs and RREPs􏸶 On the other hand􏸹 if we compare the bandwidth over􏹃 head ratio b etween the two session typ es􏸹 we 􏸷nd that the voice sessions had more optimal results than the S DATA sessions􏸶 This is b ecause for virtually the
works using the voice session tion 􏸴􏸶􏸵􏸶 We used this session
ities of AODV􏸶 The results gether with the comparable session typ e􏸹 are given in
Multicast􏸹 as a
tions􏸹 must b e considered when
gorithms for ad􏹃ho c networks􏸶
hanced AODV to provide multicast capability􏸶 Mul􏹃 ticast using AODV follows directly from the Route Request􏸿Route Reply message cycle and requires only one additional message typ e􏸹 the Multicast Validation Message􏸶 No des in the network that are memb ers of the same multicast group􏸹 together with the nodes used as routers to connect group members􏸹 form a bi􏹃 directional multicast tree across which multicast data packets are relayed􏸶 The MACT message is used to select the no de which a source no de cho oses as its next hop for the multicast tree􏸶 Additionally􏸹 there is a mul􏹃 ticast group leader that is resp onsible for incrementing the multicast group sequence numb er􏸶 More details of the multicast p ortion of AODV can b e found in 􏹆􏸵􏸾􏹇􏸶
Current
Status and
Future Work
Currently􏸹 AODV has b een
sp eci􏸷ed in an
IETF manet working
or improve resp onse time to route queries􏸹 or
increase
the capabilities of the
4.1. Multicast
proto col􏸶
to the
further improvements
basic
to ol
for conferencing designing routing al􏹃
applica􏹃 We have already en􏹃
􏸺

􏹑 of No des
􏹋􏸻
􏸵􏸻􏸻
Session Typ e
S DATA
Voice
S DATA
Voice
Go o dput Ratio at sim end
􏸾􏸺􏸶􏸽􏹋􏹖
􏸺􏹓􏸶􏸵􏸺􏹖
􏸾􏸴􏸶􏸾􏸼􏹖
􏸺􏸴􏸶􏸴􏸺􏹖
Bandwidth E􏹊ciency
􏸵􏸶􏸵􏹉
􏸵􏸶􏸻􏹓
􏸵􏸶􏸵􏸵
􏸵􏸶􏸻􏹓
Avg Rte Acq Latency 􏹌msec􏹍
􏸼􏸻􏹓
􏸴􏸺􏸺
􏸼􏸻􏸼
􏹋􏸺􏸻
􏹑 Generated Sessions
􏸼􏹉
􏹉􏹋
􏹓􏸼
􏸺􏸾
4.2. Intermediate Node Route Rebuilding
Route rebuilding after a link breakage is currently the resp onsibility of the source no de􏸶 However􏸹 one al􏹃 ternative to this metho d is to allow the no de upstream of the break to try to repair an active 􏹌i􏸶e􏸶 recently used􏹍 route before sending the link failure noti􏸷cation􏸶 Because the next hop with which it lost contact is likely to still b e in the near vicinity and have a valid route to the destination􏸹 the TTL value of the RREQ sent by the intermediate node can be small so that the link failure can b e lo calized􏸶
A tradeo􏹂 b etween quickly reestablishing the route and preventing the source no de from continuing to send data packets exists when allowing intermediate no des to rebuild routes􏸶 Allowing intermediate no de route re􏹃 building could provide for quicker route reconstruction and fewer dropp ed packets if the route is able to b e reconstructed quickly􏸶 On the other hand􏸹 more data packets will have b een lost during an unsuccessful re􏹃 construction attempt than would have b een if a link failure noti􏸷cation had b een sent at the initial discov􏹃 ery of the broken link􏸶 We plan to investigate which metho d is sup erior in terms of go o dput and latency􏸶
4.3. Elimination of Hello Messages
Hello messages􏸹 while allowing no des to learn ab out neighb or changes in a timely manner􏸹 create extra con􏹃 trol overhead and increase bandwidth consumption􏸶 We chose to include hello messages in the design of AODV b ecause we did not want AODV to have to rely on an underlying MAC􏹃sublayer proto col􏸶 However􏸹 we are currently investigating ways of eliminating the
This may􏸹 for instance􏸹 lead to transmitting additional route information along a backb one􏸶 Such backb one no des might p erform intermediate varieties of route request propagation b efore relaying such requests in􏹃 discriminately􏸹 further improving bandwidth utiliza􏹃 tion􏸶 Another alternative could b e that mobile no des that are currently corresp onding might o􏹂er to ex􏹃 change their lo cal routing tables with each other􏸹 thus reducing further the setup time required for any of their mutual neighb ors to communicate with each other􏸶
QoS is another imp ortant feature of routing proto􏹃 cols􏸶 AODV has b een enhanced to provide basic QoS services􏸹 namely delay and bandwidth assurances􏸶 We plan to investigate the e􏹊cacy of these additions in the near future􏸶
need for op erate col􏸶
hello messages􏸹 while still allowing AODV to indep endently from such an underlying proto􏹃
4.4. Locality of Association and QoS
We exp ect improvements to the latency of estab􏹃 lishing routes by exploiting lo cality of asso ciation􏸶
Table 4. Comparison of Voice and S DATA Simulations
􏸾
􏹋􏸶
Conclusion
In summary􏸹 we have presented a distance vector al􏹃 gorithm that is suitable for use with ad􏹃ho c networks􏸶 AODV avoids problems with previous prop osals 􏹌no􏹃 tably DSDV􏹍 and has the following features􏸳
􏹚 Nodes store only the routes that are needed
􏹚 Need for broadcast is minimized
􏹚 Reduces memory requirements and needless dupli􏹃
cations
􏹚 Quick resp onse to link breakage
􏹚 Lo op􏹃free routes maintained by use of destination
sequence numb ers
􏹚 Scalable to large p opulations of no des􏸶
Compared to DSDV􏸹 and other algorithms which store continuously up dated routes to all destinations in the ad􏹃ho c network􏸹 our algorithm has longer latency for route establishment􏸹 but we have taken the following steps to alleviate this problem􏸳
􏹚 A route to a destination may b e returned by any intermediate no de
􏹚 Link breakages are rep orted immediately􏸹 and routes are quickly re􏹃established
in active routes

􏹚 Inactive routes are quickly aged out of the system b ecause they are more likely to go stale􏸶
Some improvements are enabled by careful b o okkeep􏹃
􏹆􏸽􏹇 M􏸶 Gerla and J􏸶􏹃C􏸶 Tsai􏸶 Multicluster􏸹 Mobile􏸹 Mul􏹃 timedia Radio Network􏸶 ACM J􏸶 Wireless Networks􏸹 􏸵􏹌􏸴􏹍􏸹 July 􏸵􏸾􏸾􏹋􏸶
􏹆􏸺􏹇 S􏸶 Guha and S􏸶 Khuller􏸶 Approximation Algorithms for Connected Dominating Sets􏸶 University of Mary􏹃 land College Park Technical Rep ort 􏸴􏹓􏹓􏸻􏸹 June 􏸵􏸾􏸾􏹓􏸶
􏹆􏸾􏹇 J􏸶 Ioannidis and G􏸶 Q􏸶 M􏸶 Jr􏸶 The Design and Im􏹃 plementation of a Mobile Internetworking Architec􏹃 ture􏸶 In Proceedings of the Winter USENIX Confer􏹃 ence􏸹 pages 􏹉􏸾􏸵􏹙􏹋􏸻􏸼􏸹 Jan􏸶 􏸵􏸾􏸾􏸴􏸶
􏹆􏸵􏸻􏹇 D􏸶 Johnson and D􏸶 Maltz􏸶 Dynamic source routing in ad􏹃ho c wireless networks􏸶 In Computer Communica􏹃
ing and by asso ciating each route with a neighb ors􏸶
We conclude
worst􏹃case route
by the network diameter􏸹 AODV is an excellent choice for ad􏹃ho c network establishment􏸶 It will b e useful in applications for emergency services􏸹 conferncing􏸹 battle􏸷eld communications􏸹 and community􏹃based net􏹃 working􏸶 We lo ok forward to further development of the proto col for quality of service􏸹 intermediate route rebuilding􏸹 and various interconnection top ologies with 􏸷xed networks and the Internet􏸶
􏹓􏸶
Acknowledgment
􏸵􏸾􏸾􏹓􏸶
􏹆􏸵􏸵􏹇 J􏸶 Jubin and J􏸶 D􏸶 Tornow􏸶 The DARPA Packet Ra􏹃
dio Network Proto cols􏸶 In Proceedings of the IEEE􏸹
volume 􏸽􏹋􏸹 􏸵􏸹 pages 􏸼􏸵􏹙􏸴􏸼􏸹 Jan􏸶 􏸵􏸾􏸺􏸽􏸶
􏹆􏸵􏸼􏹇 B􏸶 M􏸶 Leiner􏸹 D􏸶 L􏸶 Nielson􏸹 and F􏸶 A􏸶 Tobagi􏸶 Issues
in Packet Radio Network Design􏸶 Proceedings of the IEEE Special issue on 􏹅 Packet Radio Networks􏹅􏸹 􏸽􏹋􏸹 􏸵􏸳􏹓􏹙􏸼􏸻􏸹 􏸵􏸾􏸺􏸽􏸶
􏹆􏸵􏸴􏹇 J􏸶 Macker and S􏸶 C􏸶 􏹌chairs􏹍􏸶 Mobile Ad􏹃ho c Networks 􏹌manet􏹍􏸹 􏸵􏸾􏸾􏸽􏸶 http􏸳􏸿􏸿www􏸶ietf􏸶org􏸿html􏸶charters􏸿manet􏹃 charter􏸶html􏸶
􏹆􏸵􏹉􏹇 S􏸶 Murthy and J􏸶 J􏸶 Garcia􏹃Luna􏹃Aceves􏸶 A Rout􏹃 ing Proto col for Packet Radio Networks􏸶 In 􏸵st ACM Int􏹐l Conference on Mobile Computing and Network􏹃 ing 􏹌Mobicom􏹐􏸾􏹋􏹍􏸹 pages 􏸺􏹓􏹙􏸾􏹋􏸹 􏸵􏸾􏸾􏹋􏸶
􏹆􏸵􏹋􏹇 A􏸶 Parekh􏸶 Selecting Routers in Ad􏹃Hoc Wireless Net􏹃 work􏸶 In Proceedings SBT􏸿IEEE Intl Telecommunica􏹃 tions Symposium􏸹 pages 􏹉􏸼􏸻􏹙􏹉􏸼􏹉􏸹 Aug􏸶 􏸵􏸾􏸾􏹉􏸶
􏹆􏸵􏹓􏹇 V􏸶 D􏸶 Park and M􏸶 S􏸶 Corson􏸶 A Highly Adaptive Dis􏹃 tributed Routing Algorithm for Mobile Wireless Net􏹃 works􏸶 In Proceedings of 􏸵􏸾􏸾􏸽 IEEE Conference on Computer Communications 􏹌Infocom􏹐􏸾􏸽􏹍􏸹 Apr􏸶 􏸵􏸾􏸾􏸽􏸶
􏹆􏸵􏸽􏹇 V􏸶 D􏸶 Park and M􏸶 S􏸶 Corson􏸶 A Performance Compar􏹃 ison of the Temp orally􏹃Ordered Routing Algorithm and Ideal Link􏹃State Routing􏸶 In Proceedings of IEEE International Symposium on Systems and Communi􏹃 cations􏸶 IEEE Computer So ciety Press􏸹 June 􏸵􏸾􏸾􏸺􏸶
􏹆􏸵􏸺􏹇 C􏸶 Perkins and P􏸶 Bhagwat􏸶 Routing over Multi􏹃
that􏸹
establishment latency as
within the limits
Pravin Bhagwat was an early collab orator on the de􏹃 sign of AODV􏸹 and was resp onsible for a great deal of whiteb oard illumination and appropriate course correc􏹃 tions􏸶 George Aggelou participated in the implementa􏹃 tion of some previous versions of the AODV simulation􏸹
and asked the right large p ercent of the tions􏸶
References
􏹆􏸵􏹇 A􏸶 L􏸶
questions􏸶 Steve Fullmer wrote a
co de
that was used
for the simula􏹃
Alwan􏸹 R􏸶
Kleinro ck􏸹 J􏸶 Short􏸹 􏸹 and J􏸶
Bagro dia􏸹
Commu􏹃 􏹆􏸼􏹇 R􏸶 Bagrodia and W􏸶 Liao􏸶 Maisie􏸳 A language for
mobile multimedia network􏸶 IEEE Personal
nications􏸹 􏹉􏹌􏸴􏹍􏸹 June 􏸵􏸾􏸾􏸽􏸶
design of e􏹊cient dicrete event simulation networks􏸶 IEEE Transactions on Software ing􏸹 Apr􏸶 􏸵􏸾􏸾􏹉􏸶
computer Engineer􏹃
􏹆􏸴􏹇 J􏸶 Broch􏸹 D􏸶 A􏸶 Maltz􏸹 D􏸶 B􏸶 Johnson􏸹 Y􏸶􏹃C􏸶 Hu􏸹 and J􏸶 Jetcheva􏸶 A Performance Comparison of Multi􏹃Hop Wireless Ad􏹃Ho c Network Routing Proto cols􏸶 In In Pro ceedings of the Fourth Annual ACM􏸿IEEE Inter􏹃 national Conference on Mobile Computing and Net􏹃 working􏸹 Oct􏸶 􏸵􏸾􏸾􏸺􏸶
􏹆􏹉􏹇 M􏸶 S􏸶 Corson and A􏸶 Ephremides􏸶 A Distributed Rout􏹃 ing Algorithm for Mobile Wireless Networks􏸶 ACM J􏸶 Wireless Networks􏸹 􏸵􏹌􏸵􏹍􏸹 Jan􏸶 􏸵􏸾􏸾􏹋􏸶
􏹆􏹋􏹇 S􏸶 Corson􏸹 J􏸶 Macker􏸹 and S􏸶 Batsell􏸶 Architectural Considerations
for Mobile Mesh Networking􏸹 􏸵􏸾􏸾􏹓􏸶 􏹌work in progress􏹍 http􏸳􏸿􏸿tonnant􏸶itd􏸶nrl􏸶nav y􏸶mil 􏸿mmnet􏸿mmn etRFC􏸶txt􏸶
􏹆􏹓􏹇 B􏸶 Das and V􏸶 Bharghavan􏸶 Routing in Ad􏹃ho c Net􏹃 works Using Minimum Connected Dominating Sets􏸶 In IEEE International Conference on Communications 􏹌ICC 􏹐􏸾􏸽􏹍􏸹 June 􏸵􏸾􏸾􏸽􏸶
hop Wireless Network of COMM 􏹐􏸾􏹉 􏸳 Computer 􏸼􏹉􏹌􏹉􏹍􏸳􏸼􏸴􏹉􏹙􏸼􏹉􏹉􏸹 Oct􏸶 􏸵􏸾􏸾􏹉􏸶
􏹆􏸵􏸾􏹇 C􏸶 E􏸶 Perkins and E􏸶 M􏸶 Royer􏸶 Ad Distance Vector 􏹌AODV􏹍 Routing􏸶
ao dv􏹃􏸻􏸼􏸶txt􏸹 Nov􏸶 􏸵􏸾􏸾􏸺􏸶 􏹌work in progress􏹍􏸶
N􏸶 Bamb os􏸹 Villasenor􏸶
M􏸶 Gerla􏸹 Adaptive
list of active imp osed by
determined
􏸵􏸻
tions Review 􏹙
Proceedings of SIGCOMM 􏹐􏸾􏹓􏸹 Aug􏸶
􏹆􏸼􏸻􏹇 C􏸶 Perkins􏸹 Editor􏸶 IP Oct􏸶 􏸵􏸾􏸾􏹓􏸶
􏹆􏸼􏸵􏹇 A􏸶 S􏸶 Tanenbaum􏸶 Computer Networks􏸹 􏸴rd Edition􏸹 chapter 􏹉􏸹 pages 􏸼􏹓􏸴􏹙􏸼􏹓􏹉􏸶 Prentice Hall􏸹 Englewo o d Cli􏹂s􏸹 􏸴 edition􏸹 􏸵􏸾􏸾􏹓􏸶
􏹆􏸼􏸼􏹇 F􏸶 Teraoka􏸹 Y􏸶 Yokote􏸹 and M􏸶 Tokoro􏸶 A Network Ar􏹃 chitecture Providing Host Migration Transparency􏸶 In Proceedings of the SIGCOMM 􏹐􏸾􏸵 Conference􏸳 Com􏹃 munications Architectures 􏸸 Protocols􏸹 pages 􏸼􏸻􏸾􏹙􏸼􏸼􏸻􏸹 Sept􏸶 􏸵􏸾􏸾􏸵􏸶
Mobile Computers􏸶 SIG􏹃
Communications
Review􏸹
Mobility Supp ort􏸶 RFC 􏸼􏸻􏸻􏸼􏸹
Ho c On draft􏹃ietf􏹃manet􏹃
Demand

A􏸶 Pro of of the Lo op􏹃free Prop erty
Verifying that AODV establishes only lo op􏹃free routes is easy because of the e􏹂ect of the destination se􏹃 quence numb er on the maintenance of routes􏸶 Supp ose that there is a loop in a route to a destination Z􏸹 and that no des Xi are the no des in the lo op for i􏹎􏸵􏸹􏸼􏸹􏸳 􏸳 􏸳􏸹n􏸶 As a matter of terminology􏸹 we say that Xi 􏹄points to􏹅 Xj 􏹌symbolically􏸹 Xi 􏹟 X j 􏹍 if the routing table entry of Xi for destination Z shows no de Xj as the next hop
equality would have to hold the instant the lo op was
to Z 􏸶 Then􏸹 for each Xi 􏸹 Xi 􏹟 X i􏹏􏸵 for and furthermore Xn 􏹟 X 􏸵 􏸶
i􏹎􏸵􏸹􏸼􏸹􏸳 􏸳 􏸳􏸹n􏸹
Letting k one could
Let Ti b e the destination sequence
the route entry at Xi for destination Z􏸶 Then􏸹 Ti 􏹢 Ti􏹏􏸵 whenever Xi 􏹟 Xi􏹏􏸵 􏸹 b ecause of the pro cessing of the RREP sp eci􏸷ed in Section 􏸼􏸶􏸵􏸶􏸼􏸶 Since T􏸵 􏹢 T􏸼 􏹢 􏸳􏸳􏸳 􏹢 Tn 􏹢 T􏸵􏸹 evidently the destination sequence numb ers are the same for
every if it
no de were
Xi in the routing
lo op􏸶 routing
p ossible
to create a
numb er
Moreover􏸹 lo op􏸹 this
for
􏸵􏸵
created􏸶 Furthermore􏸹
quence numb ers are all the same􏸹 the next hop infor􏹃
mation must have b een
the same RREP transmitted by the destination Z􏸶
b ecause the destination se􏹃
derived at every no de Xi from Consider now the metrics mi to the destination Z􏸶
Since Xi 􏹟 Xi􏹏􏸵 only if mi 􏹎 mi􏹏􏸵 􏹏 􏸵􏸹 mn 􏹏 􏹌n 􏹞 􏸵􏹍􏸶 But because Xn 􏹟 X􏸵 􏸹 mn and n 􏹎 􏸻􏹈 this is a contradiction􏸶
then m􏸵 􏹎 􏹎 m􏸵 􏹏 􏸵􏸹
An b e the always
just by
point to X􏸼 instead of of X􏸵 􏸶 This works
inductive argument is also p ossible􏸶
minimum length of a
construct another routing lo op of length
routing lo op􏸹
mo difying the routing table at the no de Xk to
k 􏹡 􏸴􏸹 showing that it is
a routing lo op of length 􏸼􏸶
using destination sequence
shows that routing lo ops cannot have length
routing lo ops cannot contain more than one no de􏸶 But this means that there cannot b e any routing lo ops􏸶
only p ossible to Then a simple
numb ers􏸶
RREP
handling 􏸼􏸹 so that
k 􏹞 􏸵
as long as construct argument