Cryptography Reference
In-Depth Information
In the Smartphone domain the situation is different: Devices resemble personal
computers which can establish direct connections to expensive services, such as
premium SMS services or phone numbers; thus, this threat is going to re-appear.
10.4 Future Mobile Application Security Frameworks
Although much effort was put on designing security frameworks for mobile
application environments, the number of threats and security issues is still tre-
mendously high and several principle problems remain unsolved.
The security mechanisms developed and implemented for Smartphones are often
based on technologies which were deployed for regular desktop PCs over the last
20 years. While intrusion detection systems, firewalls, mandatory access control,
signature verification, etc. have been effective in traditional computing systems,
the Smartphone domain requires new security mechanisms. These must address
newly emerging types of applications, permanent connectivity, the variety and
combination of private and public information stored on the phone, the overload
of sensor information a Smartphone can provide, the undeniable resource
constraints, and the replacement of a well-trained administrator with naïve users
lacking security expertise.
In this section we address several mechanisms that can potentially cope with the
challenges of this new runtime environment.
10.4.1 Labelling of Data
The security framework of every operating system we introduced assigns
permissions to applications or processes. Based on the permissions the secure
execution of such applications is enforced. However, what does “secure “precisely
mean in this context? It implies that the application obtains access to system
resources as expected. These resources are either APIs or direct physical resources
such as sensors, files, or maybe other processes. However, as soon as permissions
have been granted, which is often already at installation time, processes can use
interfaces or data provided at their own will. Multiple studies and various proof-
of-concept exploits have shown that this security paradigm is insufficient.
We also argue that this approach is infeasible when considering the types of
applications emerging today: We are not facing monolithic applications which are
installed once and do not change any more. Instead, we are facing mash-ups, late
binding mechanisms that integrate unknown software in the flow of information,
and we are confronted with APIs that allow inexperienced users to assemble new
application logic.
Search WWH ::
Custom Search