Abstract: Different services based on different switching
techniques in wireless networks leads to drastic changes in the
properties of network traffic. Because of these diversities in services,
network traffic is expected to undergo qualitative and quantitative
variations. Hence, assumption of traffic characteristics and the
prediction of network events become more complex for the wireless
networks. In this paper, the traffic characteristics have been studied
by collecting traces from the mobile switching centre (MSC). The
traces include initiation and termination time, originating node, home
station id, foreign station id. Traffic parameters namely, call interarrival
and holding times were estimated statistically. The results
show that call inter-arrival and distribution time in this wireless
network is heavy-tailed and follow gamma distributions. They are
asymptotically long-range dependent. It is also found that the call
holding times are best fitted with lognormal distribution. Based on
these observations, an analytical model for performance estimation is
also proposed.
Abstract: This paper presents the IP traffic analysis. The traffic
was collected from the network of Suranaree University of
Technology using the software based on the Simple Network
Management Protocol (SNMP). In particular, we analyze the
distribution of the aggregated traffic during the hours of peak load
and light load. The traffic profiles including the parameters described
the traffic distributions were derived. From the statistical analysis
applying three different methods, including the Kolmogorov Smirnov
test, Anderson Darling test, and Chi-Squared test, we found that the
IP traffic distribution is a non-normal distribution and the
distributions during the peak load and the light load are different. The
experimental study and analysis show high uncertainty of the IP
traffic.
Abstract: Encryption protects communication partners from
disclosure of their secret messages but cannot prevent traffic analysis
and the leakage of information about “who communicates with
whom". In the presence of collaborating adversaries, this linkability
of actions can danger anonymity. However, reliably providing
anonymity is crucial in many applications. Especially in contextaware
mobile business, where mobile users equipped with PDAs
request and receive services from service providers, providing
anonymous communication is mission-critical and challenging at the
same time. Firstly, the limited performance of mobile devices does
not allow for heavy use of expensive public-key operations which are
commonly used in anonymity protocols. Moreover, the demands for
security depend on the application (e.g., mobile dating vs. pizza
delivery service), but different users (e.g., a celebrity vs. a normal
person) may even require different security levels for the same
application. Considering both hardware limitations of mobile devices
and different sensitivity of users, we propose an anonymity
framework that is dynamically configurable according to user and
application preferences. Our framework is based on Chaum-s mixnet.
We explain the proposed framework, its configuration
parameters for the dynamic behavior and the algorithm to enforce
dynamic anonymity.