Figure 2(b) shows that 67% of the clients performed
430
0 200 400 600 800 1000 1200
0
2
4
6
8
10
12
Packet #
TX Rate (Mbps)
Chapter 18 Unauthorized Sensor Node Identification
rate switching, while 33% did not switch rates. Out of the nodes that did not
perform rate switching, 85% sent less than nine packets (Figure 3). If we
exclude the non-switching nodes that sent less than nine packets (assuming
that these clients were never properly authenticated to the network), then the
percent of nodes that perform rate switching becomes 92% (Figure 2(c)).
Examining only the wireless nodes that used rate switching, Figure 4 shows
that 90% transmitted more than 37 packets and 88% were connected to network
more than two minutes. Also, 85% of these nodes switched rates within
the first three minutes of their connection.
We conclude that rate switching is a phenomenon that frequently occurs.
Our results show that the longer a wireless node is a part of the network
and the more packets it transmits, the more likely rate switching is to occur.
Therefore, rate switching is a viable attribute within the RI for distinguishing
between wireless nodes. This is a result of temporal and spatial interference,
as well as the e?®ect of collisions with link layer acknowledgements [1].
7 An Approach to Sensor Node Identification
Observing the impact of the rate switching algorithm on the temporal properties
of a wireless stream is a passive way to identify sensors with di?®erent RIs.
Pages:
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680