Wireless Netw

DOI 10.1007/s11276-016-1287-y

A secure and trust based on-demand multipath routing

scheme for self-organized mobile ad-hoc networks
Gautam M. Borkar1 • A. R. Mahajan2

Ó Springer Science+Business Media New York 2016

Abstract A mobile ad hoc network (MANET) is a self- vector–secure adjacent position trust verification)  Dolphin
configurable network connected by wireless links. This Echolocation Algorithm (DEA)
type of network is only suitable for provisional commu-
nication links as it is infrastructure-less and there is no
centralized control. Providing QoS and security aware 1 Introduction
routing is a challenging task in this type of network due to
dynamic topology and limited resources. The main purpose A mobile ad hoc network (MANET) is a system of wireless
of secure and trust based on-demand multipath routing is to mobile nodes that dynamically self-organize in arbitrary
find trust based secure route from source to destination and temporary network topologies. People and vehicles can
which will satisfy two or more end to end QoS constraints. thus be internet worked in areas without a pre-existing
In this paper, the standard ad hoc on-demand multi-path communication infrastructure or when the use of such
distance vector protocol is extended as the base routing infrastructure requires wireless extension [1]. In the mobile
protocol to evaluate this model. The proposed mesh based ad hoc network, nodes can directly communicate with all
multipath routing scheme to discover all possible secure the other nodes within their radio ranges [2, 3]; whereas
paths using secure adjacent position trust verification pro- nodes that not in the direct communication range use
tocol and better link optimal path find by the Dolphin intermediate nodes to communicate with each other [4]. In
Echolocation Algorithm for efficient communication in these two situations, all the nodes that have participated in
MANET. The performance analysis and numerical results the communication automatically form a wireless network,
show that our proposed routing protocol produces better therefore this kind of wireless network can be viewed as
packet delivery ratio, reduced packet delay, reduced mobile ad hoc network [5]. Routing protocols for ad hoc
overheads and provide security against vulnerabilities and networks must deal with limitations such as high error
attacks. rates, scalability, security, quality of service, energy effi-
ciency, multicast, aggregation and node cooperation etc.
Keywords Mobile ad hoc network (MANET)  Multicast [6]. Here, qualitative properties like security and quality of
routing scheme (MRS)  Quality of service (QoS)  service are taken into account.
AOMDV–SAPTV (ad hoc on-demand multicast distance While early research effort assumed a friendly and
cooperative environment and focused on problems such as
wireless channel access and multi hop routing, security has
become a primary concern in order to provide protected
& Gautam M. Borkar communication between nodes in a potentially hostile
[email protected]
environment [7]. A MANET routing protocol is vulnerable
1
Rajiv Gandhi Institute of Technology, Versova, Andheri to many forms of attack. It may be relatively simple to
West, Mumbai, Maharashtra 400053, India watch network traffic, replay transmissions, manipulate
2
Department of Information Technology, Government packet headers, and redirect routing messages, within a
Polytechnic College, Nagpur, Maharashtra 440001, India wireless network without appropriate security provisions

123
 Wireless Netw

[8]. The primary goal of a MANET routing protocol is to the Route maintenance will process to maintain the routing
establish a correct and efficient route between a pair of in case of any link failure happened. The proposed
nodes so that messages may be delivered in a timely scheme is simulated over a large number of MANET nodes
manner. If routing can be misdirected, the entire network with wide range of mobility and the performance is eval-
can be paralyzed. Thus routing security plays an important uated. The performance of the proposed scheme is com-
role in the security of the whole network [9]. pared with the existing routing protocols.
Quality of service (QoS) is usually defined as a set of The main technical contributions of our work are sum-
service requirements that needs to be met by the network marized as follows:
while transporting a packet stream from a source to its
1. We firstly give the definition and derivation of trust,
destination [10]. QoS routing requires not only finding a
then abstract a multipath routing model, where the trust
route from a source to a destination, but a route that sat-
an entity has for an interest neighbor forms the basic
isfies the end to-end QoS requirement, in terms of band-
building block of this model. Basing on the interest
width or delay. The role of a QoS routing strategy is to
entity’s historical behaviors, multi-dimensional trust
compute paths that are suitable for different type of traffic
attributes are incorporated to reflect trust relationship’s
generated by various applications while maximizing the
complexity in various angles.
utilizations of network resources. To find a path from
2. The standard ad hoc on-demand multi-path distance
source to destination satisfying user’s requirements, to
vector protocol (AOMDV) is extended as the base
optimize network resource usage and to degrade the net-
routing protocol to evaluate the proposed secure and
work performance when unwanted things like congestion,
trust based multipath routing model. In the secure and
path breaks appear in the network [11] are the main
trust based multipath routing scheme, Hop Count,
objectives of QoS.
Secure Forward Path Trust and Secure Reverse Path
Routing is critical to QoS support, while its performance
Trust, the three metrics compose a three-dimensional
is vulnerable to changes in network topologies. In mobile
evaluation vector for routing decision and DE (Dolphin
wireless networks, such changes are mainly caused by node
Echolocation) Algorithm provide a flexible and feasi-
mobility [12]. Also security can be considered a QoS
ble route selection to establish multiple two-way
attribute. Without adequate security, unauthorized access
trusted paths without containing the untrustworthy
and usage may violate QoS negotiations. The nature of
entities instead of the shortest route.
broadcasts in wireless networks potentially results in more
3. The performance evaluation show that the proposed
security exposure [10]. The physical medium of commu-
multipath routing scheme provides better in attack
nication is inherently insecure, so we need to design
prevention and makes a development on the packets
security-aware routing algorithms for MANETs. The ulti-
delivery ratio, routing packets overhead, route discov-
mate goal of the security solutions for MANETs is to
ery frequency and intrusion detection.
provide security services, such as authentication, confi-
dentiality, integrity, anonymity, and availability, to mobile The remaining paper is organized as follows. Section 2
users. In order to achieve this goal, the security solution discusses the literature work. In Sect. 3, we describe our
should provide complete protection across the entire pro- secure adjacent position trust verification model in detail.
tocol stack [13]. Owing to the fact that traditional routing Basing on the proposed AOMDV–SAPTV routing proto-
protocols are not suitable for the unique characteristic of col, in Sect. 4, the parametric matrices analysis and
MANETs, a large number of research activities [14–26] experimental results of AOMDV–SAPTV is given. Finally,
have been carried out to explore and overcome the con- Sect. 5 gives the concluding remarks of this paper.
straints of MANETs and solve design and application
issues. The proposed multipath routing scheme (MRS)
finds stable multicast path for multimedia transmission in 2 Related work
MANET. A multipath mesh is constructed and the trans-
mission route will discover in two stages. In first stage to Paramasivan et al. [27] have used the dynamic Bayesian
maintain the quality of routing the physical parameter signaling game to analyze the strategy profile for regular
analysis will done by analyzing Transmit Energy, Distance, and malicious nodes in MANET for Routing. This game
channel load, buffer occupancy, bandwidth and bit error also revealed the best actions of individual strategies for
rate (BER). Then in second stage the security of route will each node. Perfect Bayesian equilibrium (PBE) provides a
be analyzed by using route request and route reply packets. prominent solution for signaling games to solve incomplete
One of the most stable paths with better quality for routing information by combining strategies and payoff of players
in the secure environment is discovered by employing that constitute equilibrium. This game can also furnish
Dolphin Echolocation Algorithm (DEA) technique. Then secure and reliable communication that makes effective

123
Wireless Netw

cooperation among nodes. Using PBE strategies of nodes Liu et al. [29] have proposed a new routing protocol is
are private information of regular and malicious nodes. Authenticated Anonymous Secure Routing (AASR), to
Regular nodes should be cooperative during routing and satisfy the requirement and defend the attacks. More
update their payoff, while malicious nodes take sophisti- specifically, the route request packets are authenticated by
cated risks by evaluating their risk of being identified to a group signature to defend the potential active attacks
decide when to decline. The cluster based routing protocol without unveiling the node identities. The key encrypted
(CBRP) efficiently minimizes the flooding traffic during onion routing with a route secret verification message, was
route discovery. It is suitable for a small network. In large designed to prevent intermediate nodes from inferring a
networks, it provides more overlapping cluster structures real destination and also check whether AASR can achieve
which increase the routing overhead so, they proposed ad the anonymity goals by three anonymities namely identity
hoc on demand distance vector (AODV) provides reliable anonymity, route anonymity, and location anonymity. To
data transmission in MANETs. In AODV, there was a develop the anonymous protocols, a direct method is to
requested source and destination sequence number, which anonymize the commonly used on-demand ad hoc routing
is the essential reason for the routing loop problem and for protocols, such as AODV and ANODR. These results were
privacy. This approach minimizes the utility of malicious used to compare the performance of AASR to that of
nodes and it motivates better cooperation between nodes by ANODR, in a representative on-demand anonymous rout-
using the reputation system. Regular nodes monitor con- ing protocol. The results show that, it provides more
tinuously to evaluate their neighbors using belief updating throughput than ANODR under the packet-dropping
systems of the Bayes rule. Even though the regular nodes attacks, although AASR experiences more cryptographic
are follow the PBE strategy to reduce the malicious node operation delay. Compared to ANODR, AASR provides
utilities for improving throughput in the entire networks. higher throughput and lower packets loss ratio in different
The performance analysis concludes that the PBE strategy mobile scenarios in the presence of adversary attacks. It
was the best strategy for regular nodes to reduce malicious also provides better support for the secure communications
nodes utility. In this analysis, throughput and routing that are sensitive to packet loss ratio. In future, they will
latency are about 91 % respectively, than other protocols improve AASR to reduce the packet delay. A possible
that improve the networks performance. method was to combine it with a trust based routing. With
Shen et al. [28] have proposed a QoS-Oriented Dis- the help of the trust model, the routing protocols will be
tributed routing protocol (QOD) to enhance the QoS sup- more active in detecting link failures, caused either by the
port capability of hybrid networks. Taking advantage of mobility or adversary attacks.
fewer transmission hops and any cast transmission features Qin et al. [30] have proposed a novel statistical traffic
of the hybrid networks, QOD transforms the packet routing pattern discovery system (STARS). STARS aims to derive
problem to a resource scheduling problem. QOD incorpo- the source and destination probability distribution, i.e., the
rates five algorithms: QoS-guaranteed neighbor selection probability for each node to be a message source and
algorithm to meet the transmission delay requirement, destination, and the end-to-end link probability distribu-
Distributed packet scheduling algorithm to further reduce tion, which is the probability for each pair of nodes to bean
transmission delay, A mobility-based segment resizing end-to-end communication pair. To achieve its goals,
algorithm that adaptively adjusts segment size according to STARS includes two major steps one is to Construct point-
node mobility in order to reduce transmission time, A to-point traffic matrices using the time-slicing technique,
traffic redundant elimination algorithm to increase the and then derive the end-to-end traffic matrix with a set of
transmission throughput. A data redundancy elimination traffic filtering rules, and next one is Apply a heuristic
based transmission algorithm to eliminate the redundant approach to identify the actual source and destination
data to further improve the transmission QoS. A number of nodes, and then correlate the source nodes with their cor-
queuing scheduling algorithms have proposed for Differ- responding destinations, which use the probability distri-
entiated Service (DiffServ) to further minimize packet butions produced by STARS are good indicators of the
droppings and bandwidth consumption. Analytical results actual traffic patterns, i.e., actual sources, destinations, and
based on the random way-point model and the real human end-to-end links and which reveals most of the actual end-
mobility model show that QOD can provide high QoS to end links by slightly sacrificing thefalse-positive rate.
performance in terms of overhead, transmission delay, Specifically, in most cases, more than 80 percent of the
mobility-resilience and scalability. The traffic redundant actual end-to-end links are revealed (i.e., the false-negative
elimination based transmission algorithm can further rate was less than 0.2), while the false-positive rate was not
increase the transmission throughput. In the future they more than 0.16.
placed to evaluate the performance of QOD based on the Li et al. [31] analyze the impact of network load on
real test bed. MAODV protocol, and proposed an optimized protocol

123
 Wireless Netw

MAODV-BB (Multicast Ad hoc On-demand Vector with between two physical neighbours (one-hop) are considered
Backup Branches), which improves robustness of the more reliable than those of multi-hop communications; (3)
MAODV protocol by combining advantages of the tree For the purpose of identifying misbehaving nodes, each
structure and the mesh structure. The extension of monitoring node should be equipped with some local
MAODV protocol was to construct a multicast tree with detection mechanism (4) The wireless link is symmetrical,
backup branches from two aspects. One is the process of while the ‘trust’is not necessarily symmetric between two
backup branches selection and addition, the other is the physical neighbourhood entities; (5) The cooperative
mechanism of multicast tree maintenance. It not only can action in the network interaction is encouraged, which is
update shorter tree branches but also construct a multicast naturally required in such networks.
tree with backup branches. As a tree based multicast Basing on the above assumptions, a mobile ad hoc
routing protocol, M-BB shows an excellent performance in network with n nodes can be abstracted. Due to the mobile
light weight ad hoc networks. Mathematical analysis and nodes join, leave, or fail over time, the number of n may be
this result both demonstrate that the MAODV-BB protocol dynamically changing. In such networks, trust is a rela-
improves the network performance over conventional tionship between any two physical neighbour entities,
MAODV in heavy load ad hoc networks. MAODV-BB’s which also can be described as an edge of a directed graph
packet delivery was always maintained at a high level even abstracted from the graph theory. Under permitting con-
when the network load is heavy also obvious to see that the ditions, each node in the trust system is initially authenti-
delay of MAODV-BBis always lower than MAODV’s. In cated by an authentication method. In our trust model,
MAODV-BB, the existence of backup branches reduces every node maintains a trust value for each of its neigh-
the frequency of tree reconstruction and ensures high bours. This value is a measure of the credible degree of low
packet delivery ratio in heavy load ad hoc networks. and high, defined in a continuous range between 0 and 1
(i.e., 0 B TVij B 1). Let viand vj denote the monitoring
node and the monitored node, respectively. Figure 2 shows
3 Mesh based multicast routing in mobile adhoc the Mesh based multicast routing model for Mobile Ad-hoc
network Networks (MANETs).

The group-oriented services are one of the primary appli- 3.2 Secure-trust enhanced ad-hoc on-demand multi-
cations by mobile ad hoc networks (MANETs) in recent path routing protocol (AOMDV–SAPTV)
years. To support such services, multicast routing is used.
Thus, there is a need to design stable, reliable and secured Our proposed multicast routing scheme can be incorpo-
multicast routing protocols for MANETs to ensure better rated into any routing protocol. As an application, a novel
packet delivery ratio, lower delays, reduce overheads and trust-enhanced on-demand multi-path routing protocol is
security mechanism handles misbehaviors and avoid vari-
ous attacks. To overcome the above problems occurred in
MANET, A mesh based multipath routing scheme will
proposed in this work. The process flow diagram for pro-
posed routing scheme is illustrated in Fig. 1. S D

3.1 System model assumptions

We make some assumptions before designing the secure

adjacent position trust verification framework: (1) In order
to ensure the practicality of trust model, we follow the tenet
S- Source
that the ‘trust’ should be defined and quantified locally. In D-Destination
other words, the trust value is quantified only using the
local information for scalability; (2) The communications Fig. 2 Mesh based multicast routing

Secure and Discover Multiple Two way trust Find feasible

Performance
Trust based path by AOMDV measure by optimal path by
Analysis
Route Discovery routing protocol SAPTV protocol DE Algorithm

Fig. 1 Proposed flow diagram

123
Wireless Netw

proposed (abbreviated as AOMDV–SAPTV), using the Table 1 Routing entries of AOMDV–SATPV

standard AOMDV as the base routing protocol. Any
mobile node in this trust system has the ability to quantify Destination
the trust or reliability for each neighbor and select the
trustworthy path to transmit data stream. Sequence Number
3.2.1 Routing table Broadcast Hop Count
In this paper, the path trust is determined according to the
Expiration Timeout
reliability of each node on this path. The logic is that as
soon as any node is untrustworthy, the entire path is
untrustworthy. Due to the asymmetry of ‘trust’, this vector Route List
can be classified into two types, the secure path trust and
the reverse path trust. The former is along the direction of {( NextHop1 , LastHop1 , HopCount1 ,
data flow, and the latter is the reverse direction. Precisely
speaking, the former represents the subjective judgment of SecureForwardPathTrust1 ,
the source, which is used to make a decision whether or not
to transmit the data stream by this path. And the later
represents the subjective judgment of the destination, Secure Re versePathTrust1 )
whether or not to receive the data stream from this path.
SecurePathTrust ( NextHop2 , LastHop2 , HopCount2 ,
8 9
>
> SecureForwordPathTrust ¼ minfTVmk g > >
>
> >
>
>
> s  m  d  1 >
> SecureForwardPathTrust 2 ,
>
> >
>
>
> >
>
>
< >
km þ 1= ð1Þ
¼
> Secure Re versePathTrust 2 )
>
> SecureReversePathTrust ¼ minfTVkm g >>
>
>
> >
>
>
> >
>
>
> s  m  d  1 >
> ……
>
> >
>
: ;
km þ 1

where vs is the sender, vd is the receiver, vm and vk are any ( NextHopk , LastHopk , HopCountk ,
two adjacent nodes on the candidate routing path, and
vm ? vk denotes that vk is the next hop of vm. The routing SecureForwardPathTrust k ,
table entries of proposed routing protocol AOMDV-
SAPTV can be seen in Table 1.
According to the above description, two new fields [i.e., Secure Re versePathTrust k )}
Secure Forward Path Trust (SFPT) and Secure Reverse
Path Trust (SRPT)] are added into the original routing
entries of AOMDV–SAPTV, shown in Table 2. Hop
Count, Secure Forward Path Trust and Secure Reverse
Path Trust, the three metrics compose a three-dimensional (NT). The value of field SRPT is determined based on the
evaluation vector for routing decision, which provides a minimum of the continued product of trust that the RREQ
flexible and feasible approach to establish multiple two- packet has passed on this path, which is initialized to 1 and
way trusted paths. varies with the packet transmission. We introduce NT to
represent the path trust requirement, which remaining
unchanged during this control packet flooding.
3.2.2 Secure route discovery
RREQ forwarding procedure: After an intermediate
node vj receives an RREQ packet from a neighbor nodevk,
Primarily, the source node begins a network-wide flood by
broadcasting a route request packet and waits for route Step 1 If node vjhas no route to this neighbor node vk, it
reply packets. Two new fields are added into RREQ packet, will create a route entry with the filed Secure Reverse
i.e., Secure Reverse Path Trust (IPT) and Needed Trust Path Trust (SPRTkj) = RVjkin its local routing table.

123
 Wireless Netw

Table 2 SPTA packet of AOMDV–SAPTV

0 1 2 3

0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1

Type N P Reserved Dest Count

Unreachable destination IP Address(1) Path Id(1)

Unreachable Destination Sequence Number(1)

Unreachable destination IP Address(2) Path Id(2)

Additional Unreachable destination sequence number(2)

Step 2 Then it will check whether a similar RREQ has (pathjd)]. If an intermediate node has a fresh route to the
been received or not. If so, assume both of the two destination and the received RREQ packet has not been
packets fulfill the trust requirement, while the later copy processed previously, this node will send a route reply
has no less Hop Counter or superior sequence number, (RREP) packet via reversing back the path of RREQ. If the
then the RREQ will be rejected and the process ends; destination receives multiple copies of RREQ, it will reply
otherwise, go to the next step. On the other hand, if the the first k-paths at most. The parameter k is used to control
packets unsuccessful to meet the trust requirement, they the number of RREPs and prevent an RREP storm. If an
will not be forwarded and deleted instantly. Any in- intermediate node receives an RREP, it will send the RREP
between node may receive multiple RREQ copies from via unicast unless the Secure Forward Path Trust of the
other in-between nodes, then go to step 2. route which the RREP has passed is less than the Needed
Step 3 If node vk is not the source, node vj makes a Trust. When the RREP travels back to the source via
reverse route to the source using the former hop (node traversing the path of the RREQ, each node on this path
vk) of the RREQ as its next hop. The value of filed SRPT finally can set up a trusted forwarding route.
is set to min [SPRTsk, TVjk] when TVjk is well-known,
else the value is min [SPRTsk, Trust_Value].
3.2.3 Secure path trust alert system
Step 4 If node vj has a valid route to the destination in its
routing table, and the routes Sequence Number is greater
In this paper, we propose a novel data-driven route main-
than the Dest Sequence No in the RREQ, node vj will
tenance mechanism, termed as Secure Path Trust Alert. We
generate an RREP to node vk. Otherwise, node vj
convert the route error packet into the secure path trust
modifies the SRPT of the RREQ using min [SPRTjk,
alert packet by adding a new field flag Pin the reserved
TVjk] when TVjk is known, when TVjk is unknown. Then
field after field N, shown in Table 3. The value of Flag
node vj increases the value of field Hop Count by one
P set to 1 indicates that the packet is a secure path trust
and propagates this modified RREQ packet to all
alert packet. When the path trust is lower than the trust
neighbors.
requirement value, a path trust alert event will be triggered.
Three new fields are also added into the RREP packet. The path id field could distinguish the different paths via
The field Secure Forward Path Trust (SFPT) represents the using the last hop field in its own routing table entries.
minimum of the continued product of trust that the RREP When an intermediate node receives a data packet, it
has passed in route reply, which is initialized to 1. The new will select its next hop based on the routing entry. And at
added field Needed Trust (NT) has the same meaning as the same time, it will also confirm whether its potential
that in the RREQ. And the field Secure Reverse Path Trust next hop’s trust is larger than the trust value or not. If not,
(SRPT) is set to min [SRPT (obtained in RREQ), SRPT which specifies that the next hop is not a trustworthy node

123
Wireless Netw

Table 3 Simulation parameters other trust route maintenance systems this new proposed
Parameters Details
secure path trust alert system could decrease the routing
overhead and route discovery frequency. The main reasons
Simulation tool Matlab are: (1) This new system is more worthy to the ‘trust’
Node placement Random criterion, destination, path id rather than only destination;
No. of nodes 20, 40, 60, 80, 100 (2) the earlier hop of each path in the precursor list is used
No. of sink (destination) One to govern the propagation range of the route error message.
No. of sources 100 (Node 1–100)
Area of simulation 2500 m 9 1000 m 3.2.4 Loop freedom of AOMDV–SAPTV
Packets generated by each source 250
Total packets generated in N/W 100 9 250 = 2500 On-demand protocols in MANETs may encounter routing
Size of each packet 1000 bytes loops problem due to that they discover routes with the
Initial energy 100 J help of broadcasting mechanism. Sequence number
Transmission range 250 m mechanism effectively guarantees loop freedom. That is,
Radio model Two way ground for the purpose of avoiding the possibility of any cycle,
Max speed 28 m/s each node maintains an increasing serial number. Desti-
Traffic type FTP nation sequence number are tagged on all routing packets,
MAC Mac/802_11 so as to provide a mechanism to calculate two relatively
Bandwidth 11 mb fresh routing packets generated two different nodes of the
Simulation time (in s) 1000 s same destination.
Antenna type Omni directional An intermediate node creates a reverse path to the
Link layer type LL source only when receiving a fresh control packet RREQ,
Interface queue type Queue/drop tail and a forwarding path to the destination with the RREP. At
Channel type Channel/wireless channel
some time, an intermediate node vj receives a control
Network interface type Phy/wirelesss phy
packet to a destination d (vj 6¼ d) from a neighbour vk. The
variables SequenceNumber kd, HopCounter kd and Se-
cureReversePathTrust kd represent the Dest Sequence No,
Hop Counter and Secure Reverse Path Trust of the control
packet respectively. Let SequenceNumberjd, RouteListjd,
(i.e., suspect or malicious node) and a secure path trust
MaxTrustjd and MinHopsjd be Sequence Number, Route
alert event will be activated. In this case, a SPTA packet
List, maximum Secure Path Trust and minimum Hop
will be sent to its previous hop with the help of precursor
Count of multiple paths to destination d in the routing
list in routing table via unicast.
table of node vj respectively. Combined with RREQ, the
The definitive goal of the system is in response to the
update rule for route entries in routing table is shown as
sudden or hidden malicious nodes in the routing path, in
follows.
order to maintain the efficiency of routing. Compared with

123
 Wireless Netw

Loop freedom of AOMDV-SAPTV

1. if ( SeqNumber jd < SeqNumberkd ) then //a new RREQ packet

2. SeqNumber jd = SeqNumberkd ;

3. RouteList jd =NULL;

4. if( TV jk is unknown) then New Secure Reverse Path Trust=min [ Secure Re versePathTrust kd ,

Trust_Value];

5. else if TV jk >Trust_Value, New Secure Reverse Path Trust=min[ Secure Re versePathTrust kd

, TV jk ];

6. insert (k, HopCounterk +1, New Secure Reverse Path Trust) into RouteList jd ;

7. node v j rebroadcasts the RREQ;

8. else if TV jk <Trust_Value;

9. node v j discards this packet;

10. end if

11. else if ( SeqNumber jd = SeqNumberkd then

12. if(min[ Secure Re versePathTrust kd , TV jk ]>Secure Path Trust_Value and ( HopCounterkd

<MinHopsjd-1) then

13. if( TV jk is unknown) then New Secure Reverse Path Trust=min( Secure Re versePathTrust kd ,

Trust_Value);

14. else New Secure Reverse Path Trust=min[ Secure Re versePathTrust kd , TV jk ];

15. endif

16. insert (k, HopCounterK +1, New Secure Reverse Path Trust) into RouteList jd ;

17. node v j rebroadcasts the RREQ;

18. else node v j discards this packet;

19. endif

20. endif

123
Wireless Netw

As mentioned above, line 1, 4, 8, 11, 12 and 18 of the immediately. Because of the rush, the round trip time
rule ensures loop freedom. The proposed protocol is only recorded by a route request is always smaller than the true
allowed to accept an alternate route with smaller hop count value if an attacker is on the route, and therefore the route
in accordance with meeting the trust requirement. is likely to be selected as the shortest route. AOMDV–
SAPTV uses hop count as a routing metric in Route Reply
3.3 Security analysis of AOMDV–SAPTV Phase, and thus is resistant to the rushing attack.

Since AOMDV–SAPTV can hide network topology, 3.3.4 Sybil attack

malicious nodes cannot launch attacks from central posi-
tions of the network. Thus, the potential damages incurred A sybil attacker disrupts route discovery by impersonating
by malicious nodes are greatly reduced or even eliminated. multiple legal nodes. To launch this attack, the attacker
Next, we analyse the robustness of AOMDV–SAPTV in first obtains the identity of a set of legal nodes and then
resisting the following attacks when the attacks are laun- impersonates some or all of them to participate in multiple
ched from random positions. route discoveries. AOMDV–SAPTV does not include
identity nor topology information in the routing messages,
3.3.1 Black hole attack and thus it is impossible for malicious nodes to obtain the
identity information of other nodes. Therefore, AOMDV–
A black hole attacker disrupts route discovery by forging a SAPTV is resistant to sybil attack.
route to the destination. A typical attack is launched as
follows: When Source S broadcasts a route request to 3.4 Enhanced SAPTV with adjacent secure
search a route to destination D, attacker ‘A’ replies and authentication to avoid attacks
advertises a route RAD from itself to destination D. If
source S sends packets to destination D via route RAD, the We enhance SAPTV by integrating an neighbour authen-
attacker A can intercept and discard the packets. Since tication mechanism with it. The improved SAPTV can
AOMDV–SAPTV does not allow intermediate nodes to resist more attacks. In the following, we first present the
send route reply messages, it can resist the black hole enhanced SAPTV and then analyse its robustness when
attack. facing attacks other than aforementioned ones.
Before joining MANET, every node obtains a certificate
3.3.2 Wormhole attack from a trusted certificate server T. From this certificate,
every node has a pair of public key KS0 and private key S0S ,
A typical wormhole attack is launched as follows. Two and it keeps its authentication information IMSG:
collaborating attackers first select two central positions in  
IMSG ¼ t; ID; S0S ðt; IDÞ; KS0 ;
the network to reside such that they are located on many
potential routes. Then they build a private tunnel between where ‘t’ is the lifetime of the authentication information.
them and advertise a fake hop count which is smaller than The information is used to authenticate the identity of the
the real hop count between them. The action disrupts the sender of a route message and verify the message’s
route discovery mechanisms which only use hop count as integrity. A node always inserts its IMSG into the message
routing metric since the private channel between the two it initiates.
attackers will always be selected as part of routes consid- Before we present the Route Request Phase with
ering the smaller hop account. AOMDV–SAPTV can resist neighbour authentication, we introduce the following
wormhole attack because (1) it is topology-hiding and it is notations: S and D represent the source node and the des-
impossible for attackers to choose central positions to tination node, respectively. A and B represents two inter-
launch the attack and (2) it uses round-trip time as a routing mediate nodes, and N represents a neighbour of the
metric in Route Probe Phase, which makes it robust against destination. Indicates that the message is sent via
hop count modification. broadcasting.
The RREQ packet has the following characteristics:
3.3.3 Rushing attack 1. The RREQ is anonymous
2. It is transmitted utilizing new, auto generated MAC
Rushing attack is one of the denial of service attacks. address,
While a normal node waits for a random delay before 3. It consists of a public key KS0 taken from MS0 master key
sending a packet to avoid collision in wireless communi- of anonymous single-time use keys that don’t permit
cation, a rushing attacker always forwards packets neighbors to map the key onto a particular node.

123
 Wireless Netw

We stress that retaining the identity of the verifier is

concealed. It is most essential in order to make our
AOMDV–SAPTV robust against attacks.

The Route Request Phase has multiple steps:

Step 1: S * A : RREQ = [[S , Seq, D, hopCt ]S s′ , IMSGS ]

Step2: A * B : Secure Re versePathTrust (2)

Step3: B * D : RREQ [
= [S , Seq, D, hopCt ]S s′B , IMSG B ]
The Route Reply Phase also has multiple steps:

Step 1: D *N : RREP [ ]
= [S , D, nextNode, exNodeSet ]S s′D , IMSG D

Step2: N *B : RREP = [[S , D, nextNode, exNodeSet ]S ′ , IMSG ] sN N

(3)

Step3: B * A : RREP [ ]
= [S , D, nextNode, exNodeSet ]S s′B , IMSG B

Step4: S * A : RREP = [[S , D, nextNode, exNodeSet ]S ′ , IMSG ]sS S

Neighbour Authentication message interchange protocol-Algorithm

1. Node S do

2. S → * : RREQ = [[S, Seq, D, hopC t ]S s′ , IMSGs ]

3. S: store t s

4. When receive RREP from X ⊂ N S do

5. S:store t XS , RREQ = [[S , D , nextNode , exNodeSet ]S SD

′ , IMSG D ]

6. after Tmax + Δ + Trandom do

7. Calculate Tmin

8. End

With neighbour authentication, before transmitting the transmission policy according to the mechanism in based
packets, the source node first verifies the availability of on the number of available routes. When the number of the
routes and finds the shortest secure route. Then it determines available routes is less than or equal to three, the source node

123
Wireless Netw

uses single route policy to forward packet; otherwise, mul- where i is the search space index, k is the discrete time
tiple-route policy is applied. The source node restarts the index, di is the search space dimension of ith location for a
route discovery if there is no available route. The enhanced dolphin, xi is position of ith location for a dolphin, Pi is the
AOMDV–SAPTV can resist other attacks. For example, best position found by kmwinv location for a dolphin (per-
modification attack can be detected by authenticating the sonal best), G is the best position found by dolphin (global
integrity of route messages. Impersonation attack can be best), b1i and b2i are random numbers in the interval [0, 1]
prevented because every node is required to authenticate its applied to ith location for a dolphin.
neighbours. Fabrication attack can be defeated by appending The convergence factor should change during the opti-
a signature to route messages. mization process, should be assigned. Here, the change of
CF (Convergence Factor) is considered to be according to
3.5 Find better link quality optimal path using DE the following formula:
algorithm for data transmission
IiP  1
PPðIi Þ ¼ PP þ ð1  PP1 Þ ð6Þ
An optimization technique called DEA is used to find ðIN ÞP  1
better link quality path to transfer data into our proposed PP is the predefined probability, PP1 the convergence
network scheme. DEA can be applied to optimization factor of the first iteration in which the solutions are
problems that are partially in dynamic topology changing selected randomly, Ii the number of the current loop, and
environment. DEA is applied to find the best nodes Power is the degree of the convergence curve.
involved in a path DEA is meta-heuristic that searches In our simulations, the following equation is used for
large spaces of candidate solutions. A route with a better search space dimension
link quality is selected for forwarding data from source to
destination. If a better link quality is not found, DEA di ðk þ 1Þ ¼ lðkÞdi ðkÞ þ q1 ½b1i ðPi  xi ðkÞÞ þ q2 ½b2i ðG
function is performed again until global best solution has  xi ðkÞÞ
been found. DEA reduces the traffic and routing overhead ð7Þ
of the optimization process and finds the node with best
link quality in an ad hoc network. In which lk is the inertia function q1 and q2 are the
constant factors of search speed. In this paper, linear
3.5.1 DE Algorithm for optimal route selection decreasing strategy has been used in which an initially
large inertia weight is linearly decreased to a small value as
The main steps of Dolphin Echolocation (DE) for discrete follows:
optimization are as follows: ðNI  kÞ
lk ¼ ½lð0Þ  lðNI Þ þ lðNI Þ ð8Þ
Initialize nodes (number of echolocations) in a MANET. NI
In DE (Dolphin Echolocation) algorithm, each location
where NI is the maximum number of iterations for which
for a dolphin has a search space dimension and a position
the algorithm is executed, l(0) is the initial inertia weight,
as follows:
l(NI) is the final inertia weight. Algorithm 3 describes the
di ðk þ 1Þ ¼ di ðkÞ þ b1i ðPi  xi ðkÞÞ þ b2i ðG  xi ðkÞÞ ð4Þ steps of the DEA algorithm for optimal weight vector
xi ðk þ 1Þ ¼ xi ðkÞ þ di ðk þ 1Þ ð5Þ selection to train SVM.

123
 Wireless Netw

Algorithm 3: DEA for optimal weight vector selection

Step 1: Initialize dolphin and randomize the position and search space of each location

( xi , d i ; i = 1,........., M )

Step 2: Calculate predefined probability using equation (6)

Step 3: Compute the fitness function of each location

( y (i ) = fitness( xi )).

Step 4: Calculate the accumulative fitness according to dolphin rules

Alternatives name it as A

for k = -Re to Re

1
AF( A+ k ) j = * (Re − | k |) fitness( xi ) + AF( A+ k ) j
Re

Step 5: Initialize each Pi and G as Pi 0 = y i and G = min( Pi 0 ); i = 1,......, M .

Step 6:Update the search space of location for a dolphin using dynamic inertia weight (Eqs. (8)

and (19)) and control it by search space clamping as follows:

⎧d i ( k + 1) if d i ( k + 1) < Dmax
d i ( k + 1) = ⎨
⎩ Dmax if d i ( k + 1) ≥ Dmax

Update the position of location for a dolphin (Eq. (7)).

Step 7: Update Pi and G based on the new value of fitness function as:

y i ,new = fitness ( xi ,new ), Pi = y i ,new and G = min( Pi ).

Step 8:If the stop conditions are not satisfied, go to Step 4. Otherwise, stop and return G as the

best solution.

DEA is initialized with a group of secure paths and then population. The bound of the inertial range option is use for
searches for an optimal route solution by updating genera- providing a satisfactory solution that eventually is discov-
tions. Each echolocation is updated by two best values in the ered. This best value is a global best. The DE algorithm
iterations. The first one is the best solution that has been significantly reduces the traffic overhead and computation
achieved previously. The second best value is tracked by the complexity. The DEA reduced the route failure between
dolphin rules obtained currently by any paths in the nodes that minimize the routing overhead. To decrease the

123
Wireless Netw

effect of random error, every experiment repeats 50 times load functionality was implemented by several scenarios
and the average of experimental results is used as the per- were configured for testing. This variation leads us and a
formance metrics. usefulness of a single channel load measurement. This
channel load measurement can significantly improve the
network performance both in network latency and
4 Simulation model throughput.

To configure proposed secure adjacent position trust veri- 4.1.4 Bandwidth

fication model, we used the following simulation parame-
ters which we have discussed in Table 3. Bandwidth is the rate of data transfer, bit rate or through-
put, measured in bits per second. The amount of data that
4.1 Performance metrics can be carried from one node to another in a given time
period is known as bandwidth in MANET. It measures how
To maintain the quality of routing the physical parameter much data can be sent over a specific connection in a given
analysis will done by analyzing Transmit Energy, Distance, amount of time. Now days modern networks typically have
channel load, buffer occupancy, bandwidth and bit error speeds measured in the millions of bits per second
rate (BER). (megabits per second, or Mbps) or billions of bits per
second (gigabits per second, or Gbps). However, this
4.1.1 Transmit energy estimate indicates how much bandwidth an application or
device in the wireless network can expect when sending or
We assume that the data transmission between the nodes receiving network traffic. This bandwidth variation as on
with power P the corresponding transmit energy is PTs. Let low or high frequency.
the variable Es,t be the minimal energy required to transmit Formula for the lower cutoff frequency
one data packet from the source node at (0, 0) to the des- sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi!
tination node at (D, 0), and where Z is in decibels over a 1 1
f1 ¼ f0 1þ 2 ð13Þ
path with exactly t hops 4Q 2Q
Est Formula for the upper cutoff frequency
Gt ¼ ð9Þ
Ehop sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi!
1 1
The normalized minimal transmit energy over a single f2 ¼ f0 1þ 2 ð14Þ
4Q 2Q
hop is
  n o
Formula for the Q factor
FG1 ðgÞ ¼ P Es;1  ghpo ¼ P 10z=10  g ð10Þ
    f0
Q¼ ð15Þ
10 In g f2  f1
P z  In g ¼ 1  Q ð11Þ
in10 hr
Formula for the bandwidth
f0
4.1.2 Distance f2  f1 ¼ ð16Þ
Q

The weight function is the parameter Pi,j that allows nodes where f0 is center frequency f1 is low cutoff frequency and
to select the best path. This parameter is defined by: f2 is high cut of frequency and Q is the Quality factor
Di;j Ei;j
Pi;j ¼ a  þb ð12Þ 4.1.5 Bit error rate (BER)
Tri Trj

where a and b are the weights satisfied the nodes; Di,j is the Considering a multi hop route between source and desti-
distance between node i and node j. Tri transmission range nation, the BER at the end of a link between two neigh-
of node i; Trj transmission range of node j; Ei,j is the boring nodes, denoted as BER link, depends on the signal-
maximum energy between node i and node j. to-noise ratio (SNR) at the receiving node. Finally it is
possible to show that the BER at the end of the nh-th link of
4.1.3 Channel load the multi-hop route, denoted by BERnk , can be expressed as
Y
nk

This channel load focuses on analyzing the variation of BERnk ffi 1  ½1  BERlink ðiÞ ð17Þ
i1
channel load measurements for the nodes. The channel

123
 Wireless Netw

4.1.6 Through put Figure 3 illustrates a comparison among Basic AOMDV,

AOMDV-A (AOMDV with Attack), AOMDV-AP
It is defined as the total number of packets delivered over (AOMDV with Attack Prevention), AOMDV-AP-overhead
the total simulation time. The throughput comparison change (AOMDV with Attack Prevention (changes in
shows that the three algorithms performance margins are overhead) and Improved AOMDV-AP-overhead improve-
very close under traffic load of 20 up to 100 nodes in ment (Proposed AOMDV–SAPTV) in terms of throughput
MANET scenario and have large margins when number of based on random mobility scenario by varying maximum
nodes increases to 100. number of connections (number of nodes). The numbers of
Mathematically, it can be defined as: connections were varied as 20, 40, 60, 80, 100 nodes
Throughput ¼ N=1000 respectively. At high density like from 100 numbers of
connections in Improved AOMDV-AP-overhead improve-
where N is the number of bits received successfully by all ment (Proposed AOMDV–SAPTV), the throughput increa-
destinations. ses because of packet lost is too low.
Figure 4 shows that the average end-to-end delay of
4.2 Results and analysis Basic AOMDV, AOMDV-A (AOMDV with Attack),
AOMDV-AP (AOMDV with Attack Prevention),
The parameters like throughput, transmit energy, channel AOMDV-AP-overhead change (AOMDV with Attack
load, buffer occupancy, transmit distance, bit error rate and Prevention (changes in overhead) and Improved AOMDV-
packet delivery ratio are improved as previously noted. The AP-overhead improvement (Proposed AOMDV–SAPTV).
parameter analysis of proposed routing scheme can be seen The average end-to-end delay increases with the increased
in Table 4. number of connections. The numbers of connections were

Table 4 Parameter analysis versus number of nodes

Parameters No of Throughput End to end Transmit Bit error Channel Buffer Packet
nodes (kb/s) delay (ms) energy (J) rate (%) load (%) occupancy delivery ratio
(%) (%)

Basic AODV output 20 295 20.3 1586 25.2 24.5 22.8 91.7
40 614 21.7 3388 20.8 21.9 20.9 92.3
60 925 22.6 4982 16 19.8 17.1 93.9
80 1246 23.7 6545 12.8 17.5 15 94.5
100 1561 24.6 8470 9.1 15.9 12.8 96
Basic AODV with attack 20 286 20.9 1792 27.6 27.5 26.1 90.2
40 603 22.5 3586 24.3 24.1 23.9 91.1
60 914 23.2 5187 18 22.8 21 92.8
80 1235 24.5 6864 14.7 21.1 19.2 93.6
100 1552 25.5 8779 11.9 19.9 18.1 94.9
AODV with attack prevention 20 300 20.1 1516 23.3 24.5 20.1 93
40 618 21 3246 17.5 19.2 18.6 94.5
60 935 21.9 4874 13.9 17.5 15.7 95.2
80 1254 23 6432 10.7 14.9 13.4 96
100 1571 23.8 8308 7.1 13.5 11 96.9
AODV with attack prevention but 20 305 19.2 1465 20.5 23.1 18 94.1
changes in overhead 40 624 20 3032 15.4 18.4 16.1 95.3
60 942 20.9 4710 11.2 16.1 12.9 96.1
80 1435 22.1 6310 8.9 13.9 11.2 96.9
100 1705 23 8202 6.1 11.1 9 98.1
Improved AODV with attack 20 317 18.3 1210 15.6 20.5 14.9 95
prevention but improvement in 40 637 19.1 2754 12.3 17.4 11.8 96.2
overhead
60 956 20 4453 8.8 14.2 9.7 97
80 1564 21.2 6309 6.3 11 6.8 98.2
100 1787 22 8001 5 8.3 5 99.1

123
Wireless Netw

Fig. 3 Measurement of throughput varying maximum number of Fig. 5 Measurement of transmission energy varying maximum
nodes (Kb/s) number of nodes (J)

proposed protocol AOMDV–SAPTV consumes low energy

compared to others. The life time (battery) of the node for
AOMDV–SAPTV is higher than other protocol. In the case
of a link failure, AOMDV–SAPTV has the ability to make
longer battery and node’s life time because of the proper
utilization in choosing a path.
Figure 6 shows Bit Error Rate of Basic AOMDV,
AOMDV-A (AOMDV with Attack), AOMDV-AP
(AOMDV with Attack Prevention), AOMDV-AP-overhead
change (AOMDV with Attack Prevention (changes in
overhead) and Improved AOMDV-AP-overhead improve-
ment (Proposed AOMDV–SAPTV). Above graph com-
parison shows Bit Error rate is too low for our proposed
AOMDV–SAPTV protocol because of high low packet
loss.

Fig. 4 Measurement of end to end delay varying maximum number

of nodes (ms)

varied as 20, 40, 60, 80, 100 nodes. After increasing

number of connections more than 40, end-to-end delay
increase much higher because of queuing and retransmis-
sion delay. In heavy traffics load as the maximum number
of connections increase, the number of packets delivery
also increase. But based on the above graph comparison
end to end delay for our proposed AOMDV–SAPTV is
very low.
Figure 5 shows transmission energy of Basic AOMDV,
AOMDV-A (AOMDV with Attack), AOMDV-AP
(AOMDV with Attack Prevention), AOMDV-AP-overhead
change (AOMDV with Attack Prevention (changes in
overhead) and Improved AOMDV-AP-overhead improve-
ment (Proposed AOMDV–SAPTV) and the maximum
number of connections energy consumption respectively. Fig. 6 Measurement of bit error rate varying maximum number of
Based on the above graph comparison shows that our nodes

123
 Wireless Netw

Fig. 7 Measurement of channel load varying maximum number of Fig. 8 Measurement of buffer occupancy varying maximum number
nodes of nodes

Figure 7 shows channel load percentage of Basic

AOMDV, AOMDV-A (AOMDV with Attack), AOMDV-
AP (AOMDV with Attack Prevention), AOMDV-AP-
overhead change (AOMDV with Attack Prevention
(changes in overhead) and Improved AOMDV-AP-over-
head improvement (Proposed AOMDV–SAPTV). Above
graph comparison shows channel load percentage is too
low for our proposed AOMDV–SAPTV protocol because
of traffic occurrence level is very low.
Figure 8 indicates the effect of buffer occupancy
ofBasic AOMDV, AOMDV-A (AOMDV with Attack),
AOMDV-AP (AOMDV with Attack Prevention),
AOMDV-AP-overhead change (AOMDV with Attack
Prevention (changes in overhead) and Improved AOMDV-
AP-overhead improvement (Proposed AOMDV–SAPTV).
Above graph shows the proposed routing protocol Fig. 9 Measurement of packet delivery ratio varying maximum
AOMDV-SAPV using the multipath but congestion number of nodes
avoiding ability of proposed protocol gives better
throughput then the AOMDV. AOMDV–SAPTV uses the
buffer space of the neighboring node so packet drop is less distance, bit error rate and packet delivery ratio are improved
as compared to the AOMDV. So it shows that AOMDV– compared to other existing routing protocols. It denotes our
SAPTV is better than AOMDV. proposed AOMDV–SAPTV provides better quality of ser-
Figure 9 shows packet delivery ratio of Basic AOMDV, vice (QoS) and security against vulnerabilities.
AOMDV-A (AOMDV with Attack), AOMDV-AP
(AOMDV with Attack Prevention), AOMDV-AP-overhead
change (AOMDV with Attack Prevention (changes in 5 Conclusion
overhead) and Improved AOMDV-AP-overhead improve-
ment (Proposed AOMDV–SAPTV). Above graph com- Mobile ad hoc networks have attracted much interest in the
parison shows PDR rate is too high for our proposed research community due to their potential applications.
AOMDV–SAPTV protocol because of secure trust based However, the inherent characteristics of such networks
optimal route selection. make them vulnerable to a wide variety of attacks. The
Based on the above parametric matrices like throughput, security concerned in these wireless networks remains a
transmit energy, channel load, buffer occupancy, transmit serious impediment to widespread adoption. In this paper,

123
Wireless Netw

we focus on the security of routing protocol in MANETs. MANETs. IEEE Journal on Selected Areas in Communications,
Firstly, we abstract a secure adjacent position trust verifi- 22(7), 1347–1356.
13. Wan, J., Liu, J., Shao, Z., Vasilakos, A. V., Imran, M., & Zhou,
cation model. Then by extending the standard ad hoc on- K. (2016). Mobile crowd sensing for traffic prediction in internet
demand multi-path distance vector protocol (AOMDV), we of vehicles. Sensors, 16(1), 88.
propose a novel secure adjacent trust-enhanced routing 14. Dvir, A., & Vasilakos, A. V. (2011). Backpressure-based routing
protocol combined with the trust model, named as protocol for DTNs. ACM SIGCOMM Computer Communication
Review, 41(4), 405–406.
AOMDV–SAPTV. The persuasive experiments have been 15. Zhang, X. M., Zhang, Y., Yan, F., & Vasilakos, A. V. (2015).
conducted to simulate and present the effectiveness of this Interference-based topology control algorithm for delay-con-
new protocol. The main purpose of QoS aware routing is to strained mobile ad hoc networks. Mobile Computing, IEEE
find a feasible path from source to destination which will Transactions, 14(4), 742–754.
16. Vasilakos, A. V., Zhang, Y., & Spyropoulos, T. (Eds.). (2011).
satisfy two or more end to end QoS constrains. The DE Delay tolerant networks: Protocols and applications. Boca
algorithm is used to find the optimal and best path for Raton: CRC Press.
routing. The proposed scheme is compared to the existing 17. Vasilakos, A. V., Li, Z., Simon, G., & You, W. (2015). Infor-
routing protocols. The result shows that our proposed mation centric network: Research challenges and opportunities.
Journal of Network and Computer Applications, 30(52), 1.
technique enhanced the quality of routing and had find the 18. Yao, G., Bi, J., & Vasilakos, A. V. (2015). Passive IP traceback:
best path by the optimization algorithm. Disclosing the locations of IP spoofers from path backscatter.
Information Forensics and Security. IEEE Transactions, 10(3),
471–484.
19. Yan, Z., Zhang, P., & Vasilakos, A. V. (2014). A survey on trust
References management for Internet of Things. Journal of network and
computer applications, 42, 120–134.
1. Corson, M. S., Macker, P. J., & Cirincione, G. H. (1999). Inter- 20. Yang, H., Zhang, Y., Zhou, Y., Fu, X., Liu, H., & Vasilakos, A.
net-based mobile ad hoc networking. Internet Computing, IEEE, V. (2014). Provably secure three-party authenticated key agree-
3(4), 63–70. ment protocol using smart cards. Computer Networks, 58, 29–38.
2. Attar, A., Tang, H., Vasilakos, A. V., Yu, F. R., & Leung, V. 21. Liu, B., Bi, J., & Vasilakos, A. V. (2014). Toward incentivizing
(2012). A survey of security challenges in cognitive radio net- anti-spoofing deployment. Information Forensics and Security,
works: Solutions and future research directions. Proceedings of IEEE Transactions, 3, 436–450.
the IEEE, 12(100), 3172–3186. 22. Zhou, J., Cao, Z., Dong, X., Xiong, N., & Vasilakos, A. V. (2015).
3. Cordasco, J., & Wetzel, S. (2008). Cryptographic versus trust- 4S: A secure and privacy-preserving key management scheme for
based methods for MANET routing security. Electronic Notes in cloud-assisted wireless body area network in m-healthcare social
Theoretical Computer Science, 197(2), 131–140. networks. Information Sciences, 314, 255–276.
4. Azedine, B., El-Khatiba, K., Xua, L., & Korbab, L. (2005). An 23. Wei, L., Zhu, H., Cao, Z., Jia, W., & Vasilakos, A. V. (2010).
efficient secure distributed anonymous routing protocol for Seccloud: Bridging secure storage and computation in cloud. In
mobile and wireless ad hoc networks. Computer Communica- 2010 IEEE 30th international conference distributed computing
tions, 28(10), 1193–1203. systems workshops (ICDCSW) (pp. 52–61).
5. Li, Wenjia, & Joshi, Anupam. (2008). Security issues in mobile 24. Wang, T., Liu, Y., & Vasilakos, A. V. (2015). Survey on channel
ad hoc networks—A survey (pp. 1–23). Baltimore County: reciprocity based key establishment techniques for wireless sys-
Department of Computer Science and Electrical Engineering, tems. Wireless Networks, 21(6), 1835–1846.
University of Maryland. 25. He, D., Chen, C., Chan, S., Bu, J., & Vasilakos, A. V. (2012). Re
6. Jing, Q., Vasilakos, A. V., Wan, J., Lu, J., & Qiu, D. (2014). Trust: Attack-resistant and lightweight trust management for
Security of the internet of things: Perspectives and challenges. medical sensor networks. Information Technology in Biomedi-
Wireless Networks, 20(8), 2481–2501. cine, IEEE Transactions, 16(4), 623–632.
7. Yang, Hao, HaiyunLuo, Fan Ye, Songwu, Lu, & Zhang, Lixia. 26. Zhou, J., Dong, X., Cao, Z., & Vasilakos, A. V. (2015). Secure
(2004). Security in mobile ad hoc networks: Challenges and and privacy preserving protocol for cloud-based vehicular DTNs.
solutions. Wireless Communications, IEEE, 11(1), 38–47. Information Forensics and Security, IEEE Transactions, 10(6),
8. Wei, L., Zhu, H., Cao, Z., Jia, W., & Vasilakos, A. V. (2010). 1299–1314.
Seccloud: Bridging secure storage and computation in cloud. In 27. Paramasivan, B., Prakash, M. J. V., & Kaliappan, M. (2015).
IEEE 30th International Conference on 2010 distributed com- Development of a secure routing protocol using game theory
puting systems workshops (ICDCSW) (pp. 52–61). model in mobile ad hoc networks. IEEE Journal of Communi-
9. Deng, H., Li, W., & Agrawal, D. P. (2002). Routing security in cations and Networks, 17(1), 75–83.
wireless ad hoc networks. Communications Magazine, IEEE, 28. Shen, H., & Li, Z. (2014). A QoS-oriented distributed routing
40(10), 70–75. protocol for hybrid wireless networks. IEEE Transactions on
10. Mohapatra, P., Li, J., & Gui, C. (2003). QoS in mobile ad hoc Mobile Computing, 13(3), 693–708.
networks. IEEE Wireless Communications, 10(3), 44–53. 29. Liu, W., & Yu, M. (2014). AASR: Authenticated anonymous
11. Wu, C., Zhang, F., & Yang, H. (2010). A novel QoS multipath secure routing for MANETs in adversarial environments. IEEE
path routing in MANET. International Journal of Digital Content Transactions on Vehicular Technology, 63(9), 4585–4593.
Technology and its Applications, 4(3), 132–136. 30. Qin, Y., Huang, D., & Bing, Li. (2014). STARS: A statistical
12. Jiang, S., Liu, Y., Jiang, Y., & Yin, Q. (2004). Provisioning of traffic pattern discovery system for MANETs. IEEE Transactions
adaptability to variable topologies for routing schemes in on Dependable and Secure Computing, 11(2), 181–192.

123
 Wireless Netw

31. Li, X., Liu, T., Liu, Y., & Tang, Y. (2014). Optimized multicast Dr. A. R. Mahajan is working
routing algorithm based on tree structure in MANETs. China as Head, Department of Infor-
Communications, 11(2), 90–99. mation Technology, Govern-
ment Polytechnic College,
Nagpur, India. She has obtained
Gautam M. Borkar received her Ph.D. in Computer Science
his Bachelors degree from and Engineering. She has pub-
National Institute of Technol- lished twenty four papers in
ogy, Jalandhar, Punjab, India international journals and one in
and completed his masters from national journal. She has pre-
Sant Gadge Baba Amravati sented forty three and five
University, Amravati Presently papers in international confer-
he is working as Assistant Pro- ences and national conferences,
fessor in Rajiv Gandhi Institute respectively. She has more than
of Technology, Mumbai and 20 years of teaching and
pursuing Ph.D. from Sant Gadge research experience. Her area of specialization is compiler opti-
Baba Amravati University, mization, artificial intelligence, parallel algorithms. She is a member
Amravati His current research of IEEE, ISTE and CSI.
interest includes network secu-
rity, trust management and
security in wireless sensor network.

123