英语论文网

m
jammers save energy by sleeping, they are less effective.Our contribution is to develop jamming attacks that (1)
work on encrypted packets, (2) are as effective as constant/deceptive/reactive jamming, and (3) at the same time more
energy-efficient than random jamming or reactive jamming.We implement such jamming attacks by exploiting the semantics
of the data link layer and show the results quantitatively.The fact that our attacks are applicable to three
representative MAC protocols suggests the same attacks areapplicable to a wide range of other protocols belonging to
the same categories as these protocols’. Our analysis of theattacks provides new insights into the timing considerationsof MAC protocols with regards to security, and provides
hints on which category of protocols provides the best protectionagainst our attacks so far, in the absence of effective
countermeasures. The motivation of this work stems fromthe concern that if an attacker can program and deploy a
general-purpose link-layer jamming network that is able tojam any WSN effectively and energy-efficiently, and if a high
entry barrier is not maintained for such a low cost attack, aWSN can never in any practical sense be secure. A counterargument
might be that energy efficiency is no concern to
powerful attackers, but even powerful jammers come with a
finite energy supply and they would advertise their presence
and location if they simply blast away with an exorbitant
amount of radio waves – this is something a sensible attacker
would avoid.
The paper is organized as follows. We start by stating the
assumptions on which our attacks are based in Section 2. We
then describe the attack algorithms in Section 3. Section 4
describes how the protocols and the corresponding attacks
are simulated, and how the results are evaluated. The results
are given in Section 6. Implications of our work to
other protocols are discussed in Section 7. Section 8 explores
some potential countermeasures. Related work is discussed
in Section 9. Finally Section 10 concludes.
2. ASSUMPTIONS
We assume an attacker has two goals: the primary goal is
to disrupt the network by preventing messages from arriving
at the sink node, and the secondary goal is to increase
the energy wastage of the sensors. A sink node is a node
that requests for, and hence sinks, information. Our attacks
depend on three assumptions: (1) the jammer motes
know the preamble used by the victim nodes, (2) the jammer
motes can measure the length of a packet, and (3) the
jammer motes know what MAC protocol that victim notes
are running. A preamble is a bit sequence, usually consisting
of alternating 1’s and 0’s, for training the receiver, and
its length depends on the data coding scheme used [34]. Requirement
(1) and (2) should be easy to satisfy in practice.
Requirement (3) is more demanding but not impractical to
satisfy – as our future work, a strategy will be devised to
map observed traffic to specific classes of protocols. Note
that the jammer motes do not need to know the content
of the packets, so our attacks work even if the packets are
encrypted. Adding to the significance of our attacks is that
the attacker does not need to capture and compromise any
existing sensor nodes.
Jammer motes
Sensor nodes
Interceptor
informs motes
of estimated
frequencies
Figure 1: Distri