Review for Secure Data Aggregation in Wireless Sensor Networks
Trust-based Security for Mobile Ad-hoc Networks using Small World Phenomenon
1. Trust-based Security for Mobile Ad-hoc
Networks using Small World Phenomenon
Abdeen M.R.1
, Fernando S.S.N.2
, Gunaratne G.Y.C.L.3
, Mallawa Arachchi T.L.4
,
Rupasinghe P.L.,5
Senaratne A.N.6
Sri Lanka Institute of Information Technology, Malabe, Sri Lanka
1. rizanabdeen@gmail.com, 2. shavinka_fernando@yahoo.com, 3.
yehan_gunaratne@hotmail.com, 4. tharuka1@hotmail.com, 5. lakmal.r@sliit.lk, 6.
amila.n@sliit.lk
Abstract
Mobile Ad-hoc Networks (MANETs) are rapidly
growing in popularity. A collection of mobile
devices (nodes) connected in an ad-hoc manner for
data transferring purposes is called a MANET.
These networks have high performance, but they
lack proper security features which can guarantee
the confidentiality, integrity and availability of
data. Any security mechanism for MANETs should
take into consideration the limited power resources
of mobile devices. Hence, cryptographic solutions,
with their complex calculations do not provide
energy efficient security. Trust-based security
however can overcome this limitation to a certain
extent. This research paper mainly focuses on
building a trust algorithm that can generate a
quantitative value for trust with the help of the
Small World Phenomenon.
KEY WORDS
Mobile Ad-hoc Network, MANET
Security, Trust, Small World
Phenomenon
1. Introduction
MANET is rapidly becoming an
important topic in the field of
communication. It promises to be an
essential feature in the not too distant
future. MANET is a group of mobile
nodes, which forms an impermanent
network without the support of
centralized administration or standard
support services regularly available on
conventional networks. MANETs are
featured by dynamic topology
(infrastructure-less), multi-hop
communication, limited resources
(bandwidth, CPU, battery, etc.) and
limited security [1], [2].
In its current state, any device can
connect to a MANET and start
communicating without too many
restrictions. This brings about a great
threat to the security (Confidentiality,
Integrity and Availability) of the
MANET. Many security solutions have
been proposed by many researchers.
However, a cryptographic approach will
not be feasible as the mobile devices have
limited power and therefore cannot
repeatedly execute complex
cryptographic calculations. Also, this
method will impact the performance of a
MANET. The key is to maintain the
delicate balance between performance
and security.
The Trust approach however, does not
impact performance in a significant way.
The problem with trust is that it is a
qualitative term and it is relative to each
individual. In our research, we convert
trust into a quantitative value in order to
rate the trustworthiness of a mobile
device. Stanley Milgram’s experiment
yielded that it only took an average of 4
people to connect any two individuals in
the world. This is the Small World
Phenomenon.
2. Research Methodology
2.1. Overview
The main goal is to create an algorithm to
define trust in a quantitative manner. The
basic outline of the methodology is as
2. follows. Each device will need to create a
unique ID using its IMEI number and
SIM number. Then each device will have
to calculate its trustworthiness using the
algorithm. This value cannot be edited.
Consider two devices X and Y. If X needs
to connect to Y, it will display its unique
ID to Y. If it has connected with Y
recently, Y’s unique ID will be in the
cache of X. Then it can connect to Y. If
not in the cache, it can request another
neighbour if they have Y’s ID in their
cache. Failing which, X can view Y’s
trust level and decide whether to connect
to Y or not.
2.2. Algorithm Development
Antenna Power
PRx = PTx* A
4*π*r2
Since A, 4 and π are constants, we can
say that,
PTx α PRx
r2
PRx– Power Received
PTx – Power Transmitted
A – Area of antenna
r – Distance between nodes
For a trusted node, we can say that,
PKTA = Packets received per unit time
No. of neighbor nodes
So for any node,
PKTN = Packers received per unit time
No. of neighbor nodes
PKTN Can be < or > than 1
PKTA
Multiply this by 10
Weight = 40%
Value based on Trusted List
10 – Connecting node already in the
trusted list
06 – Matching node exists in both lists
03 – No matching node but has been
previously connected to a MANET
Weight = 40%
Phone Rating
Phone rating is also used here with a
weight of 20%
Final Trust Value =
PTx[
2PKTN
PKTA
+ R+2V]
5r2
Algorithm: A node joining with
another node in the MANET
Figure 1: Algorithm for a node
connecting to another node
Algorithm: Phone Rating
Battery
The battery level (amount of power left)
is checked initially to see if this phone
can be used in a MANET or not. If the
battery level is below 10% the phone
rating will be equal to 1.0 implying that
the phone cannot be used in a MANET
else phone rating will be calculated.
3. CPU
Calculate Pi (π) up to 100 decimal points
and repeat it 500 times. Then measure the
time taken.
>500ms → 2
400ms – 499ms → 4
300ms – 399ms → 6
200ms – 299ms → 8
<200ms →10
Weight = 3
RAM (units in MB)
Get the available free memory of the
RAM and according to the available
memory allocate a score.
512 – 800 → 3
801 – 1024 → 5
1025 – 1500 → 8
> 1500 → 10
Weight = 5
Applications
Identify the apps which are harmful and
allocate a score in regard to how the
phone is safe.
(
𝑇𝑜𝑡𝑎𝑙 𝐴𝑝𝑝𝑠−𝐻𝑎𝑟𝑚𝑓𝑢𝑙 𝐴𝑝𝑝𝑠
𝑇𝑜𝑡𝑎𝑙 𝐴𝑝𝑝𝑠
) × 10 × 4
Weight = 4
Ports
Identify the open and closed ports and
according to the result allocate a score.
(
𝑈𝑛𝑢𝑠𝑒𝑑 𝑃𝑜𝑟𝑡𝑠
𝑇𝑜𝑡𝑎𝑙 𝑃𝑜𝑟𝑡𝑠
) × 10 × 2
Weight = 2
Data
3G → 10
EDGE → 6.7
GPRS → 3.4
Weight = 2
PHONE RATING = ∑ (
𝑾𝒊×𝒊𝑺𝒄𝒐𝒓𝒆
𝟏𝟔
)𝒏
𝒊=𝟏
3. Research Findings
Parameters to determine the trust value
have been implemented in the proposed
algorithm. Due to testing purposes,
certain dynamic parameters had to be
made static in order to run the simulation.
Packet ratio – We cannot obtain real
values since this is a simulated
environment.
Phone rating – Had to be hardcoded since
parameters such as CPU, battery life etc
are unavailable in the simulation
environment.
The scenario is simulated with ten nodes
for several times and the results can be
explained as below.
The trust values obtained vary from less
than 1 to values near 900. Approximately
90% of the times, the values were less
than 1. The critical trust value was altered
several times and the optimum critical
value should be a value less than 1. It is
possible to assign a value larger than this
for the critical value though. However,
we should keep the critical value low to
preserve the smooth data transmission
over MANET. It operates much securely
with very less effect to performance.
Further tests should be carried out using
other dynamic parameters (Packet ratio
and phone rating). So the dynamic nature
of the trust value should be changed more
rapidly according to security condition. In
this case, the new critical value must be
determined by observing trust values
closely.
4. Conclusion and Future Work
Our goal was to provide a security
solution for MANETs. As mentioned in
our proposal, the plan was to implement
the algorithm on real devices. Currently,
the best protocol for MANET routing is
BATMAN. After studying it well, we
4. realized that it does not suit our
requirements.
AODV was selected due to its low
complexity. We came across a few issues
since there was no actual implementation
in the real world. Hence, we too followed
suit.
Due to limitations in the simulation
environment, we had to set certain
dynamic parameters as static ones. We
managed to obtain trust values
accordingly, which was a major
milestone. At the same time, we
developed an Android application to
generate a unique ID for each phone in a
MANET and to calculate a rating for the
phone.
We were however, unable to implement
the protocol in real devices. The results
we obtained are promising and it will
definitely open the path for more research
in this field. This algorithm can be further
developed to accommodate real devices.
References
[1] Piyush Patidar ”Mobile Ad-Hoc
Network (MANETs)”,Swan Jain
Academy, Indore, 2007
[2] P. Sinha, R. Sivakumar, and V.
Bharghavan, âA˘ IJCedar: Core
extraction distributed ad hoc routing,
â˘A˙Iin Proc. of IEEE INFOCOM,
1999.