InternetPerils delivers Intelligence Streams in a variety
of forms for several types of customers, using multiple sources,
data fusion, and deep analysis with pattern recognition to reveal
the big picture of Internet anomalies. The most distinctive features of the
core technologies are intertwined among data collection, analysis,
and visualization, enabling pattern recognition at all three
Data collection via a traceroute-like probing mechanism enables
direct, constantly updated, simultaneous observation of Internet topology
and Internet performance. The basic patterns of
topology are inherent in the collected data.
Availability of combined topology and performance data permits
automated selection of critical nodes, both baseline and during
perils. This form of pattern recognition enables both presentation
of results and tuning of focus sets for further results.
Selection of important nodes permits abstraction, making it
possible to visualize important features of massive datasets.
Visualization presents automated results of analysis, permits
manual investigation and interpretation of those results.
Visualization enables manual pattern recognition, which
facilitates tuning analysis and data collection.
InternetPerils' specific graphing, selection,
and probing techniques enable several
advantages at the three levels of collection, analysis, and
visualization. Most importantly, their use together provides the
lightweight, automated, yet manually tunable methods which produce
the frequent, ongoing, detailed representation of patterns of
InternetPerils enables companies to manage Internet business risk.
InternetPerils' Gap Analysis of Internet Networks (GAIN) creates a
living topology of the Internet, examining various depths, different
parts, and different functions of the Internet. InternetPerils
uses all-source data fusion to understand the complete picture, by
integrating data from hundreds of feeds from dozens of sources.
This must be done
constantly with a full understanding of the interrelationships among the
pieces. InternetPerils collects its own Internet measurements from sensors
and combines it with news feeds and human intelligence. The result is a
current, constantly-updated, map of the entire fabric and framework of
Then InternetPerils uses deep analysis to make sense of the data by
pattern recognition and event detection, in order to identify weak points
in the Internet, recommend redundancy strategies as well as conduct
survivability modeling, bottleneck and peering failure analysis. This
combination of all-source data fusion and deep analysis provides a
timely, holistic, synoptic view of the Internet. For insurance customers,
InternetPerils aggregates this Internet Intelligence into tables
for actuarial use. InternetPerils also provides visualizations of
perils and anomalies for use by insurance brokers, security assessors,
PerilScope makes devices and data pathways, as well as their
performance qualities, as immediately apparent as a railway diagram.
It can analyze and display logical topologies and performance of the
Internet from any number of abstraction levels, revealing perils as
small as a nonredundant link and anomalies as large as cyber-hurricanes.
PerilScope can select subjects of interest from the cloud, which can be
used to build actuarial matrices based on performance histories over time.
Fused with other data, such as information about ownership and operational
specifics of nodes, a quantitative picture of very high resolution can
be drawn of a topological space in which commercial data is resident.
At the high levels of abstraction PerilScope can provide the kind of
structured performance intelligence that gives risk managers insight
into the global aggregated risks that should inform the pricing of all
Internet business risk management products.