probability of accurate detection. We expect this will prove
more efficient. We will also run actual experiments with
different data models on a large campus network and perform
trace-driven simulations to characterize interesting scenarios.
Additionally, we plan to create the appropriate function
that will take the observed statistics as input and generate the
probability of Rogue AP Index (RAI) parameter. Once the
statistics and probability rogue function has been determined
and deemed acceptable, we will extend our analysis to
support non-traditional (other than first in first out (FIFO))
queuing at switches (i.e., queues that have priority associated
with them - an example would be an intranet that supported
voice over IP (VoIP), where voice traffic has higher priority
than data traffic in the LAN).
Our final area of interest deals with the ability to perform
the rogue AP identification in an automated fashion.
Specifically, this functionality should be a utility that can run
on a switch. Our current scheme uses a visual approach that
would prove challenging for a computer system to analyze.
We are looking at an approach where we can use the area
under the curves of the traffic shown in Figures 3-7 to
automate the analysis. As observed in the figures, the area
under the wired curves is significantly greater than that under
the wireless curves. Thus, by computing and comparing the
areas, we can potentially perform this rogue AP
identification, using inter-packet spacing, without human
intervention.
probability of accurate detection. We expect this will prove
more efficient. We will also run actual experiments with
different data models on a large campus network and perform
trace-driven simulations to characterize interesting scenarios.
Additionally, we plan to create the appropriate function
that will take the observed statistics as input and generate the
probability of Rogue AP Index (RAI) parameter. Once the
statistics and probability rogue function has been determined
and deemed acceptable, we will extend our analysis to
support non-traditional (other than first in first out (FIFO))
queuing at switches (i.e., queues that have priority associated
with them - an example would be an intranet that supported
voice over IP (VoIP), where voice traffic has higher priority
than data traffic in the LAN).
Our final area of interest deals with the ability to perform
the rogue AP identification in an automated fashion.
Specifically, this functionality should be a utility that can run
on a switch. Our current scheme uses a visual approach that
would prove challenging for a computer system to analyze.
We are looking at an approach where we can use the area
under the curves of the traffic shown in Figures 3-7 to
automate the analysis. As observed in the figures, the area
under the wired curves is significantly greater than that under
the wireless curves. Thus, by computing and comparing the
areas, we can potentially perform this rogue AP
identification, using inter-packet spacing, without human
intervention.
การแปล กรุณารอสักครู่..
