This approach is well suited for sensor networks because it does not inject
tra?±c, does not require client software, operates independent of higher layer
protocols and can work in the presence of encrypted tra?±c.
Given a trace of wireless tra?±c in which the rate switching algorithm has
been invoked, we apply Fourier analysis to create a spectrum. The spectral
features capture the temporal behavior of the trace influenced by the rate
switching algorithm. We distinguish between RIs by comparing their spectral
profiles.
7.1 Encoding Trace into Signal
Before we can apply spectral analysis, we must encode the tra?±c capture
into a signal, a time series of events. Even with an encrypted payload, the
802.11 header o?®ers a rich source of information, such as packet size, duration
of payload, packet type, and retransmitted packets. Additionally, the tra?±ccapturing
utility records a time stamp with each packet as it is collected. This
information can be used to construct various types of signals.
Our goal is to construct a packet arrival timeseries from the wireless traf-
fic stream. First we extract only data frames as indicated by the Subtype
fields within the header. This step isolates the actual data transmission frames
from other overhead communication to minimize the amount of noise in the
frequency domain. Next, we use a sampling bin of s seconds and define the
arrival process, x(t), as the number of data packets that arrive in the bin[t,
t+s].
Pages:
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680