def getBwDifferential(self):
self.iat_length_tuples = [l for l in self.iat_length_tuples if len(l) > 0]
if time.time() - self.session.startTime > 0.2 and len(self.iat_length_tuples) > 1:
lengths0 = [t[1] for t in self.iat_length_tuples[-2]]
lengths1 = [t[1] for t in self.iat_length_tuples[-1]]
bw0 = sum(lengths0) / self._period
bw1 = sum(lengths1) / self._period
bw_diff = (bw1 - bw0) / self._period
self.session.bw_diffs.append(bw_diff)
log.debug("[bwdiff %s] - bw diffs: %s", self.end, self.session.bw_diffs)
log.debug("[bwdiff %s] - abs bwdiff (%s) > threshold (%s)",
self.end, abs(bw_diff), self._threshold)
if abs(bw_diff) > self._threshold:
# we should sample uniformly from the passed iats
# convert self.iat_length_tuples to distribution
# and pass it to wfpad iat distribution as dict.
iats = gu.get_iats([t[0] for t in gu.flatten_list(self.iat_length_tuples)])
h = {iat: 1 for iat in iats}
self._burstHistoProbdist['snd'] = hs.new(h)
self._gapHistoProbdist['snd'] = hs.new(h)
else:
self._burstHistoProbdist['snd'] = hs.uniform(const.INF_LABEL)
self._gapHistoProbdist['snd'] = hs.uniform(const.INF_LABEL)
self.iat_length_tuples.append([])
log.debug("[bwdiff %s] A period has passed: %s", self.end, self.iat_length_tuples[-3:-1])
if self.isVisiting():
log.debug("[bwdiff %s] Calling next period (visiting = %s, padding = %s)",
self.end, self.isVisiting(), self.session.is_padding)
cm.deferLater(self._period, self.getBwDifferential)
def getBwDifferential(self):
self.iat_length_tuples = [
l for l in self.iat_length_tuples if len(l) > 0
]
if time.time() - self.session.startTime > 0.2 and len(
self.iat_length_tuples) > 1:
lengths0 = [t[1] for t in self.iat_length_tuples[-2]]
lengths1 = [t[1] for t in self.iat_length_tuples[-1]]
bw0 = sum(lengths0) / self._period
bw1 = sum(lengths1) / self._period
bw_diff = (bw1 - bw0) / self._period
self.session.bw_diffs.append(bw_diff)
log.debug("[bwdiff %s] - bw diffs: %s", self.end,
self.session.bw_diffs)
log.debug("[bwdiff %s] - abs bwdiff (%s) > threshold (%s)",
self.end, abs(bw_diff), self._threshold)
if abs(bw_diff) > self._threshold:
# we should sample uniformly from the passed iats
# convert self.iat_length_tuples to distribution
# and pass it to wfpad iat distribution as dict.
iats = gu.get_iats(
[t[0] for t in gu.flatten_list(self.iat_length_tuples)])
h = {iat: 1 for iat in iats}
self._burstHistoProbdist['snd'] = hs.new(h)
self._gapHistoProbdist['snd'] = hs.new(h)
else:
self._burstHistoProbdist['snd'] = hs.uniform(const.INF_LABEL)
self._gapHistoProbdist['snd'] = hs.uniform(const.INF_LABEL)
self.iat_length_tuples.append([])
log.debug("[bwdiff %s] A period has passed: %s", self.end,
self.iat_length_tuples[-3:-1])
if self.isVisiting():
log.debug(
"[bwdiff %s] Calling next period (visiting = %s, padding = %s)",
self.end, self.isVisiting(), self.session.is_padding)
cm.deferLater(self._period, self.getBwDifferential)
def relayGapHistogram(self, histo, removeTokens=False, interpolate=True,
when="rcv", decay_by=0):
"""Specify histogram that encodes the delay distribution
The delay distribution represents the probability of sending a
single additional padding packet after a given delay following
a padding packet that originated at this hop. In both cases, the
padding packet is sent in the direction of the client.
Parameters
----------
histo : list
Contains delay distribution of sending an IGNORE cell after
sending an IGNORE cell.
labels : list
Millisecond labels for the bins (with "infinity" bin to allow
encoding the probability of not sending any padding packet in
response to this packet).
removeTokens : bool
If True, follow Adaptive Padding token removal rules.
If False, histograms are immutable.
interpolate : bool
If True, randomize the delay uniformly between bin labels
If False, use bin labels as exact delay values.
when : str
If "rcv", this histogram applies to locally-inserted padding
packets that were initially sent in response to client-originated
data. If "snd", this histogram applies to packets sent in response
to locally-inserted padding packets sent in response to upstream
data. Note that this means that implementations must maintain this
metadata as internal state as the system transitions from
BURST_HISTOGRAM initiated padding into GAP_HISTOGRAM initiated
padding.
decay_by: int
To add to the "Infinity" bin after each successive padding packet
is sent. Used to create an increasing likelihood of hitting the
termination condition with each successive padding packet.
"""
histo = cast_dictionary_to_type(histo, float)
self._gapHistoProbdist[when] = hist.new(histo,
interpolate=bool(interpolate),
removeTokens=bool(removeTokens),
decay_by=decay_by)
self._deferGapCallback[when] = self._gapHistoProbdist[when].removeToken
def relayBurstHistogram(self, histo, removeTokens=False, interpolate=True,
when="rcv", decay_by=0):
"""Specify histogram encoding the delay distribution.
The delay distribution represents the probability of sending a single
padding packet after a given delay in response to either an upstream
cell, or a client-originating cell.
Parameters
----------
histo : list
Contains delay distribution of sending an IGNORE cell after
sending an IGNORE cell.
labels : list
Millisecond labels for the bins (with "infinity" bin to allow
encoding the probability of not sending any padding packet in
response to this packet).
removeTokens : bool
If True, follow Adaptive Padding token removal rules.
If False, histograms are immutable.
interpolate : bool
If True, randomize the delay uniformly between bin labels
If False, use bin labels as exact delay values.
when : str
If set to "rcv", this histogram governs the probability of
sending a padding packet after some delay in response to a packet
originating from the other PT end. If set to "snd", this histogram
governs padding packets that are transmitted after a packet
arrives from upstream. In both cases, the padding packet is
sent in the direction of the client.
"""
histo = cast_dictionary_to_type(histo, float)
self._burstHistoProbdist[when] = hist.new(histo,
interpolate=bool(interpolate),
removeTokens=bool(removeTokens))
self._deferBurstCallback[when] = self._burstHistoProbdist[when].removeToken
def test_refill_histo(self):
h = histo.new({1: 1})
h.removeToken(1)
self.assertEqual(h.hist[1], 1)
def test_remove_tokens_with_only_positive_on_right_with_inf_label(self):
h = histo.new(TEST_DICT_INF, removeTokens=True)
h.removeToken(0.1)
h.removeToken(0.1)
self.assertEqual(h.hist[0.2], h.template[0.2] - 1)
def test_remove_tokens_with_gaps_with_inf_label(self):
h = histo.new(TEST_DICT_INF, removeTokens=True)
h.removeToken(0.5)
h.removeToken(0.5)
self.assertEqual(h.hist[0.4], h.template[0.4] -1)
def test_remove_tokens_with_interpolation_with_inf_label(self):
h = histo.new(TEST_DICT_INF, removeTokens=True)
h.removeToken(const.INF_LABEL)
self.assertEqual(h.hist[const.INF_LABEL],
h.template[const.INF_LABEL] - 1)
def test_remove_tokens_with_interpolation(self):
h = histo.new(TEST_DICTIONARY, removeTokens=True)
x = h.randomSample()
h.removeToken(x)
label = h.getLabelFromFloat(x)
self.assertEqual(h.hist[label], h.template[label] - 1)
def test_remove_tokens_without_interpolation(self):
h = histo.new(TEST_DICTIONARY, interpolate=False, removeTokens=True)
x = h.randomSample()
h.removeToken(x)
self.assertEqual(h.hist[x], h.template[x] - 1)
def test_mean_of_histogram(self):
h = histo.new(TEST_DICTIONARY, interpolate=False)
self.assert_mean(h, 0.3)
