Lakhasly

In this paper we present an attacker-centric threat model for mobile platforms.

In recent years the expanding mobile hand-held device
market is becoming an increasingly attractive target for
malicious attacks. According to recent security reports
[1][2], the number of possible malicious exploits and
executed attacks is going to surge in 2011. This trend can
be attributed to two key factors: the ever increasing user
base and the emergence of smart-phone technology. The
size of mobile device market is clearly visible from the
latest reports issued by the ITU [3] which indicate that by
the end of the 2010 there will be an estimated 5.3 billion
mobile phone users in the world. Although malicious
exploits for mobile phones have been steadily developing
over the last decade [4][5], the constraints in both
hardware and operating systems have limited the attacks
both in their scale and impact. Therefore, the endorsement
of smart-phone technology, which provides more
computing power and functionality, is proving to be a
turning point in development of malicious exploits for
mobile hand-held devices. The estimates indicate that the
market share of smart-phones in the US could exceed 50%
of the total mobile hand-held device market by the end of
2011 [6]. Since the rest of the developed world is showing
similar trends, a significant increase in smart-phone user
base is expected over the next few years, thus making this
platform an attractive target for malicious attacks. In
addition, the smart-phones are starting to resemble PCs
both in their capabilities and the way people use them.
The smart-phone mobile platforms, like Android, iOS,
Blackberry OS, Symbian and Windows Mobile are
increasingly resembling PC operating systems[7]. Thus,
the standard malicious attacks for PCs, like worms and
trojans, as well as attack vectors, like the Internet access,
are becoming applicable to the mobile platforms. Since
the modern mobile platforms can be installed on devices
other than smart-phones, like tablets or other appliances,
the same security issues exist. We will shortly refer to all
devices with a mobile platform installed as mobile devices
in the rest of this paper.
Although rich in basic features, mobile platforms can
be extended by installing applications in much the same
way it is possible for PC operating systems. Since
applications can originate from third party providers, they
present an opportunity for introduction of malicious
exploits. Apart from utilizing computing power provided
by mobile devices, the attackers are starting to target the
data. This is due to the fact that the smart-phones are
becoming storage units for personal information through
use of various social networking applications, personal
organizers and e-mail clients.
In this paper we present an attacker-centric threat
model for mobile platforms. The threat model addresses
three key issues of mobile device security: attacker's
goals, attack vectors and mobile malware. First, it defines
the motives for attacking mobile platforms in order to
identify the attacker‟s interests and potential targets. The
attack goals focus on the motives introduced by modern
mobile platforms and devices. Second, the model
incorporates the attack vectors in order to present possible
entry points for malicious content on mobile devices.
Finally, the model considers threat types applicable to
mobile platforms if the presented attack vectors are
successfully utilized.
Since we consider the mobile applications to be the
most convenient method for performing malicious attacks,
we analyze the security model implemented by two
widely spread platforms: the Android by Google and the
iOS by Apple. The analyzed mobile platforms differ in
their approaches to enforcing application security. The
Android isolates applications in order to prevent them
from interfering with other applications or the operating
system. On the other hand, the iOS applications are
screened for malicious intentions by code reviewers, thus
allowing implementation of simpler security mechanisms.
In order to demonstrate how Android‟s permission based
security model could be breached, we present a fictive
mobile application.