Module olivia.networkmetrics
Olivia network vulnerability metrics.
Expand source code
"""Olivia network vulnerability metrics."""
from olivia.packagemetrics import Reach
def failure_vulnerability(olivia_model, metric=Reach, normalize=False):
"""
Compute network vulnerability to failure metric.
Vulnerability to failure is the mean of some cost metric for the packages in the network, measuring
the expected cost caused by the failure of an uniformly chosen random package.
For example, it corresponds to the expected number of packages (using the Reach metric)
or dependencies (using the Impact metric) potentially compromised by the introduction of a defect into a random
package.
If normalized, the result is given in terms of proportion to the total network cost (i.e. number of packages or
dependencies for the Reach/Impact metrics)
Parameters
----------
olivia_model: OliviaNetwork
Input network.
metric: Cost metric, optional
Usually Reach or Impact, but a custom metric class may be provided, assuming it implements
a compute() method that returns a MetricStats() object.
normalize: bool, optional
Normalize the output value.
Returns
-------
vulnerability: float
Vulnerability to failure value.
"""
ms = olivia_model.get_metric(metric)
if normalize:
return ms.values.mean() / ms.normalize_factor
else:
return ms.values.mean()
def attack_vulnerability(olivia_model, metric=Reach, normalize=False):
"""
Compute network vulnerability to attack metric.
Vulnerability to attack is the maximum of some cost metric for the packages in the network, measuring
the potential cost caused by a directed attack to the network.
For example, it corresponds to the maximum number of packages (using the Reach metric)
or dependencies (using the Impact metric) potentially compromised by the introduction of a defect into a package
selected by a possible attacker.
If normalized, the result is given in terms of proportion to the total network cost (i.e. number of packages or
dependencies for the Reach/Impact metrics).
Parameters
----------
olivia_model: OliviaNetwork
Input network.
metric: Cost metric, optional
Usually Reach or Impact, but a custom metric class may be provided, assuming it implements
a compute() method that returns a MetricStats() object.
normalize: bool, optional
Normalize the output value.
Returns
-------
vulnerability: float
Vulnerability to attack value.
"""
ms = olivia_model.get_metric(metric)
if normalize:
return ms.top()[0][1] / ms.normalize_factor
else:
return ms.top()[0][1]Functions
def attack_vulnerability(olivia_model, metric=olivia.packagemetrics.Reach, normalize=False)-
Compute network vulnerability to attack metric.
Vulnerability to attack is the maximum of some cost metric for the packages in the network, measuring the potential cost caused by a directed attack to the network. For example, it corresponds to the maximum number of packages (using the Reach metric) or dependencies (using the Impact metric) potentially compromised by the introduction of a defect into a package selected by a possible attacker.
If normalized, the result is given in terms of proportion to the total network cost (i.e. number of packages or dependencies for the Reach/Impact metrics).
Parameters
olivia_model:OliviaNetwork- Input network.
metric:Cost metric, optional- Usually Reach or Impact, but a custom metric class may be provided, assuming it implements a compute() method that returns a MetricStats() object.
normalize:bool, optional- Normalize the output value.
Returns
vulnerability:float- Vulnerability to attack value.
Expand source code
def attack_vulnerability(olivia_model, metric=Reach, normalize=False): """ Compute network vulnerability to attack metric. Vulnerability to attack is the maximum of some cost metric for the packages in the network, measuring the potential cost caused by a directed attack to the network. For example, it corresponds to the maximum number of packages (using the Reach metric) or dependencies (using the Impact metric) potentially compromised by the introduction of a defect into a package selected by a possible attacker. If normalized, the result is given in terms of proportion to the total network cost (i.e. number of packages or dependencies for the Reach/Impact metrics). Parameters ---------- olivia_model: OliviaNetwork Input network. metric: Cost metric, optional Usually Reach or Impact, but a custom metric class may be provided, assuming it implements a compute() method that returns a MetricStats() object. normalize: bool, optional Normalize the output value. Returns ------- vulnerability: float Vulnerability to attack value. """ ms = olivia_model.get_metric(metric) if normalize: return ms.top()[0][1] / ms.normalize_factor else: return ms.top()[0][1] def failure_vulnerability(olivia_model, metric=olivia.packagemetrics.Reach, normalize=False)-
Compute network vulnerability to failure metric.
Vulnerability to failure is the mean of some cost metric for the packages in the network, measuring the expected cost caused by the failure of an uniformly chosen random package. For example, it corresponds to the expected number of packages (using the Reach metric) or dependencies (using the Impact metric) potentially compromised by the introduction of a defect into a random package.
If normalized, the result is given in terms of proportion to the total network cost (i.e. number of packages or dependencies for the Reach/Impact metrics)
Parameters
olivia_model:OliviaNetwork- Input network.
metric:Cost metric, optional- Usually Reach or Impact, but a custom metric class may be provided, assuming it implements a compute() method that returns a MetricStats() object.
normalize:bool, optional- Normalize the output value.
Returns
vulnerability:float- Vulnerability to failure value.
Expand source code
def failure_vulnerability(olivia_model, metric=Reach, normalize=False): """ Compute network vulnerability to failure metric. Vulnerability to failure is the mean of some cost metric for the packages in the network, measuring the expected cost caused by the failure of an uniformly chosen random package. For example, it corresponds to the expected number of packages (using the Reach metric) or dependencies (using the Impact metric) potentially compromised by the introduction of a defect into a random package. If normalized, the result is given in terms of proportion to the total network cost (i.e. number of packages or dependencies for the Reach/Impact metrics) Parameters ---------- olivia_model: OliviaNetwork Input network. metric: Cost metric, optional Usually Reach or Impact, but a custom metric class may be provided, assuming it implements a compute() method that returns a MetricStats() object. normalize: bool, optional Normalize the output value. Returns ------- vulnerability: float Vulnerability to failure value. """ ms = olivia_model.get_metric(metric) if normalize: return ms.values.mean() / ms.normalize_factor else: return ms.values.mean()