Name: Katalyst

Text: Building your OSINT
Capability

T

W

A

R

White Paper

S

O

F

T

W

A

R

E

Kapow Katalyst for OSINT
Harvest text in any language, images, audio, video from websites, blogs, and social media.
Secure and non-attributa ble. Kapow Katalyst-the best-kept secret in OSINT.
I

------

HARVEST
A N OSlNT
~
~

A T A
WITH

KATAIv
-

OSlNT data sources are as varied as the lnternet itself.
Mission-critical data can reside in blogs, in news feeds, in
social media-and can even be hosted on short-lived sites
on the dark web. As technology standards continue to
evolve one thing is certain: OSlNT data sources will
remain a moving target-in more ways than one.
Only Kapow KatalystTM
has the deep understanding of
Web page structure needed to handle any site, including
those built with advanced AJAX techniques. Without that
capability, data cannot be completely or accurately
harvested from websites-and that's not good enough
when national security is at stake.

CAPTUREDATAANONYMOUSLY
Kapow Katalyst supports the widest range of options for
non-attribution. Examine and alter JavaScript on any page
before it is run. Automatically vary wait intervals,
crawling schedule, and clicks per session. Alter HTTP
headers to report different operating systems and
browsers. Integrate with anonymization solutions. Alter
proxy server in response to blocking attempts. Finally,
don't draw attention to your mission by registering for
API access-Katalyst can work with or without APls.
FederalSales@KapowSoftware.com

EXTRACTDATAI N ANY LANGUAGE
Katalyst offers built-in support for multi-byte character
encodings such as Chinese and bidirectional languages
such as Arabic and Hebrew. Katalyst is in daily use
throughout the IC to harvest the contents of news sites,
blogs, RSS feeds, and social media around the world.

"FFLINE
Katalyst enables offline retention of any web page,
domain, or list of domains for ongoing study and archival
purposes. A Kapow Snapshot is a complete, fully static
copy of any targeted site that includes all images,
documents, and other linked assets. You can access it at
any time without running the risk that the site owner will
be alerted to your activity. Only Kapow offers both-a
static Snapshot, and access to live websites.
'

:ORCE ALL

SECUR~NREQUIREMENTS

Kapow Katalyst provides the widest range of security
options-that's why it's in use on SIPRNET, JWICS, and
other secure networks. Katalyst is stateless, so it can
safely reside in your DMZ: there is no risk of
compromising data because Katalyst retains no data.
Contact us to discuss other deployable security measures.

www.KapowSoftware.com

02011, Kapow Software. All Rights Reserved.

PERFORM
BROADAND SURGICALCRAWLS
OSlNT projects use different crawl techniques at different
stages of an investigation but most approaches and COTS
products cannot accommodate this requirement. In the
early stages a "broad" crawl is often used-but other
solutions cannot scale sufficiently to support the high
data volume that Katalyst can manage. Kapow Katalyst
users routinely access thousands of web sites in a session
with linear scaling. In later stages "surgical" crawls will
seek to develop in-depth information on a narrow subject
area. This type of crawl requires advanced website
navigation and data transformation abilities and produces
a much richer data set.
Surgical crawls can become quite sophisticated as
analysis requirements become better understood. For
example, it is often necessary t o provide context and
semantic richness as soon as an extracted item is
recognized, rather than in a downstream process. For
example, by recognizing a known terrorist during a broad
crawl, a surgical crawl could be triggered to extract
additional contextual data from internal and external
sources. Then, by providing metadata and data facets
based on context available only during the crawlcontext unavailable later-downstream analysis tools can
create timelines, scatter plots, heat maps, relationship
graphs, and other advanced data representations t o
further assist analysis. Kapow customers routinely use
Katalyst to perform this kind of access.

Harvesting Deep Web data is much more complex than
mere link-following, and Kapow Katalyst offers unique
capabilities that enable Deep Web access. With its ability
to navigate complex menu structures and generate input
for web-based forms, Kapow Katalyst can automatically
navigate arbitrary menu levels and issue multiple queries
to extract entire databases and prepare them for use in
any format necessary.
Data can be extracted from within content management
systems and behind Web-based database query formsthen formatted for analyst use and automatically kept
current with Katalyst's de-duplication. Other extraction
technology, limited to following specific links that already
exist, are incapable of harvesting Deep Web data.

MAINTAIN
Mlccln~
RFADINESS
Kapow Katalyst uses patented technology that is resilient
t o page format changes, helping maintain readiness for
any data capture your mission requires
~ D KAPOW
D
TOYOUR

OSINT TEAM

These are just a few of the capabilities that set Kapow
Katalyst apart from other OSlNT extraction technology.
Contact us today to learn more about why we're so
widely deployed in the IC. Now and in the future, you'll
be pleased when you decide to make Kapow Katalyst part
of your OSlNT team.

Kapow Software may be the best-kept secret in OSINT, but
we're easy t o work with.

Kapow Software is the leader in OSlNT Data Extraction and
Application Integration.

We're GSA-listed and carry clearances.

Kapow KatalystTMautomatically extracts, transforms,
integrates and migrates data fromvirtually any source on the
web, in the cloud or across the enterprise, tovirtually
anywhere, including mobile devices, without requiring APls.

We focus solely on OSINT Data Collection-and we're good at
it--so we work well with integrators and all the technology
vendors in your OSlNT environment.
We're used daily throughout the IC t o harvest data for Link
Analysis and Visualization, Entity Extraction, Natural
Language Processing, Search, Indexing.
For OSINT information, please call us at 703.489.1445 or email
FederalSales@KapowSoftware.com

FederaISales@KapowSoftware.com

Nearly 500 customers, including AT&T, Audi, Intel, Morgan
Stanley, Vodafone, and dozens of federal agencies rely on
Kapow for data syndication, mobile-enablement, cloud and
web app integration, content migration, business
intelligence, terrorism tracking, comparison pricing - and
much more.

www.KapowSoftware.com

02011, Kapow Software. All Rights Reserved.

Contents
Executive Summary .......................................................................................................................................1
OSINT: Architecture Overview ......................................................................................................................2
Harvesting Data from the lnternet ......................................................................................................2
Modern Web Page Architecture ............................................................................................................... 3
Data Harvesting for OSlNT ............................................................................................................................ 4
Broad vs . Surgical Crawl ............................................................................................................................ 4
Extensible Data Extraction Support .......................................................................................................... 4
Non-Attribution.........................................................................................................................................5
Resilience to change .................................................................................................................................
5
Flexible crawl strategy ..........................................................................................................................5
Multiple language support ........................................................................................................................ 6
Security ..................................................................................................................................................... 6
Automated data de-duplication ................................................................................................................

7

Offline Snapshot ....................................................................................................................................... 7
Advanced Data Normalization ..................................................................................................................7
Deep Web Access ................................................................................................................................7
Data Integration for OSlNT ...........................................................................................................................8
Broad-Based Support for Standards ......................................................................................................... 8
Near Real-time Access ..............................................................................................................................

8

Partner Ecosystem .................................................................................................................................... 9
Mission Support for OSINT......................................................................................................................... 10
Operational Agility ............................................................................................................................. 10
10
Scalability ................................................................................................................................................
Production Support .................................................................................................................................11
Build Your OSlNT Capability with Kapow ................................................................................................... 12

02011. Kapow Software. All Rights Reserved

Executive Summary
Intelligence agencies across the Federal government are turning to OSlNT for the additional breadth and
depth of information that it promises t o deliver. For example, DOD Instruction 3115.12 directs DOD t o
"pursue the full integration and exploitation of open sources for intelligence purposes" and Intelligence
Community Directive (ICD) 200 states that "Analysts must be able t o find, access and share information,
analytic judgments and each other in any media and in real time."
However, unlike typical IT systems which can accept only a limited range of input, OSlNT data sources
are as varied as the lnternet itself-and will continue to evolve as technology standards evolve.
This white paper will alert you t o the key issues that can derail any OSlNT project. We will focus on the
three major functional areas relevant t o any OSlNT capability:
Data Harvesting: this is the "front end" of any OSlNT process. Raw data is accessed, extracted

from the irrelevant content that surrounds it, enriched with attribution that is available only as
data is being harvested, and prepared for downstream use
Data Integration: in this part of an OSlNT process, raw data is integrated, enhanced, and

enriched according t o the requirements of the specific mission
Mission Support: after an OSlNT capability has been developed, additional functionality will be
needed t o support and enhance the capability as mission parameters change
This white paper will also illustrate the advantages that have led over 500 customers-including
defense, intelligence, law enforcement, and other Federal agencies-to select Kapow as a key
component of their OSlNT data collection and acquisition strategies. With Kapow Katalyst these
organizations discreetly capture more data faster, and respond more quickly to changing conditions.

02011,Kapow Software. All Rights Reserved

OSINT: Architecture Overview
Just as each agency has a unique mission,
each supporting OSINT initiative has unique
requirements. However, after years of

Enrich

experience the OSINT community has
0

begun t o develop a set of best practices,
coalescing around architecture broadly

Web Data
Enhanc
0 Y

similar t o the illustration on the right.
In this highly abstracted view, an initial step captures raw internet data and prepares it for a series of
subsequent processing steps. These subsequent steps carry out a wide variety of mission-specific
processing on the harvested data, enhancing and enriching it, and ultimately preparing a corpus for
analysis. Finally, intelligence analysts access the corpus to carry out their work with additional missionspecific tools. In most cases a feedback loop is used when interim results expose the need to further
augment the corpus by acquiring and enhancing additional raw data in a new cycle.

Harvesting Data from the lnternet
Although the above abstraction applies
t o virtually all intelligence gathering and
analysis efforts, OSINT practitioners
focus on harnessing the breadth and
depth of data available on the Internetmore specifically, on accessible websites
both inside and outside their agency's
firewall-and making this data
meaningful and available to analysts.

Q

Web Data
Enhance

Kapow
Extraction
Browser

KapoW
Partner
Ecosystem

Defense and
Intelligence
Communities

Extracting data from web pages,
however, presents a large number of
significant technical challenges that must be resolved before regular operational use can be made of
Web-sourced data.
For example, extracting data from pages that use "Web 2.0" techniques-techniques that dynamically
construct web page content and structure on-the-fly, as a page is being rendered-demands deep
understanding of browser and JavaScript technology, and requires a protracted investment to ensure
that evolving standards and browser technology are supported in a timely manner.
Kapow Software has the kind of deep understanding that can be earned only through a decade of
experience, supported by continuous investment in browser-based data extraction. Kapow's unique
Extraction Browser is in daily use across the defense and intelligence communities, extracting a wide
variety of data from websites around the world and preparing it for further enrichment by members of
our partner ecosystem.

02011, Kapow Software. All Rights Reserved
S

O

F

T

W

A

R

E

Modern Web Page Architecture
The very complexity of a modern Web page forms a significant barrier t o automated data extraction.
Any given Web page is composed of a large number of elements with varying syntax and semanticsand because Web standards continue to evolve, continuous effort is needed t o stay current.
Early Web pages were coded only with HTML, which was used only t o specify how data was to be
formatted for display. Each user request resulted in a complete unit of HTML being sent to the browser.
However, as web standards have evolved, more advanced "dynamic" techniques have replaced static
web page construction, presenting much greater technical challenges.
Dynamic techniques, known
collectively as AJAX (Asynchronous
contained on
Client Web
Web Page as
Page
Third-Party
JavaScript and XML), enable a page t o
Viewed
Components
Servers
be built under control of one or more
JavaScript programs that run in the
browser client as the page is loading.
These JavaScript programs can
perform a wide variety of actionsthey can alter page format, request
database access, send and receive
messages, load media files or
documents, and even cause other
JavaScript programs to be retrieved
and executed. In addition, JavaScript is commonly used to capture web page analytics and select userspecific advertising to be displayed along with the requested page content. Pages built with AJAX
routinely involve multiple servers-not only the server hosting the selected Web page.
Some solutions use the Microsoft Windows HTTP Services (WinHTTP) API, but this has serious
limitations. The API itself does not provide access t o all of the page-creation activity necessary to
correctly extract web data, and operationally it forces crawlers t o run on Windows.
In contrast, Kapow's Extraction Browser-used t o support your crawlers as they run-contains a full
JavaScript engine that enables Kapow Katalyst t o see all dynamic activity as a web page is built, and
Kapow also handles cookies, session data, authentication data, and other browsing artifacts that allow
significantly more complete data extraction . One critical advantage that Kapow 'S JavaScript engine
provides is the ability for you to alter the flow on any web page in real time-an advantage that
improves privacy and security as raw data is being captured. Finally, Kapow Katalyst supports Windows
and several UNlX and Linux variations, giving you the widest range of deployment options.

02011, Kapow Software. All Rights Reserved

Data Harvesting for OSlNT
To better understand why Kapow has become the web data extraction tool of choice this section will
examine each of the capabilities needed to harvest web data from a web site and prepare it for
downstream enrichment and enhancement.

Broad vs. Surgical Crawl
Most OSlNT efforts take the approach of casting a wide net during initial stages of investigation-often
called a "broad crawlu-and focusing more sharply with a subsequent "surgical crawl" t o fill important
gaps in knowledge. Most automated approaches and COTS products fail because they are unable to
accommodate both requirements.
A broad crawl will generate a high volume of data but most solutions cannot scale sufficiently t o support
high data volume. Kapow Katalyst is unique in this regard; it offers multi-server configurations that
provide linear scalability capable of supporting the largest of crawls. Kapow customers routinely access
thousands of web sites in a completely automated manner without experiencing bottlenecks.
Surgical crawls, often developed later in a project, typically require a different set of capabilities. For
example, extracting data from news sites or blogs will require the ability to navigate through a series of
menus, enter search criteria, scroll through a variable number of web pages, and extract only the latest
data from a variable number of entries on each page. This enhances data quality and removes noise that
diminishes the quality of subsequent analysis.
Surgical crawls can become quite sophisticated as analysis requirements become better understood. For
example, it may be desirable to provide greater context and semantic richness for specific extracted
items that are recognized, much as Google includes a map of Milton Keynes among the search results
when "MK6 4HH" is entered. By recognizing a known terrorist during a broad crawl, a surgical crawl
could be triggered t o extract additional contextual data from internal and external sources. Then, by
providing metadata and data facets based on context available only during the crawl-context which is
not available later-downstream analysis tools can create timelines, scatter plots, heat maps,
relationship graphs, and other advanced data representations that can further assist analysis. Kapow
customers routinely use Katalyst t o perform this kind of access.

Extensible Data Extraction Support
Because it's not possible t o predict what type of data will need t o be extracted, both broad and surgical
crawls will need the widest possible data extraction functionality. Katalyst includes the ability to extract
any type of data from any Web site, including text, Office documents, PDF, flat files, images, audio and
video files, and even files accessible via FTP or other non-HTTP protocols.
A wide range of transformation capabilities can be applied during initial data extraction or any time
before data is delivered for further enrichment and enhancement, giving you the widest possible
latitude in designing a flexible extraction and delivery strategy. As one example of the sophistication
available, one customer that harvests academic papers from websites uses Katalyst to extract the

02011,Kapow Software. All Rights Reserved
S

O

F

T

W

A

R

E

abstract from each PDF-formatted paper during the crawl-no other software is needed, and no code
needed to be written. Many other similarly powerful functions are available for use, saving development
time and improving crawl performance.
If you require specialized extraction or transformation capabilities beyond those provided with Kapow,
you can invoke Java or .NET applications, as well as Web services, as you are crawling a site. Because you
can create any specialized data extraction you wish, you can increase the quality of extracted data
earlier in the process, improving responsiveness and moving closer to real-time extraction.

Non-Attribution
OSlNT collection often requires repeated visits t o a website, and it may become necessary to avoid
being noticed. Kapow Katalyst supports the widest range of options for shielding your identity.
You can examine and alter the execution of any JavaScript code on a page before it is run, enabling a
broad range of techniques for protecting identity such as defeating the use of JavaScript intended t o
track visitors. Kapow Katalyst also provides a wide range of ways to alter the flow of a crawler to make it
more human-like-by inserting dynamic wait intervals, by crawling at different times of day, and by
limiting the number of clicks in a crawl. Kapow Katalyst can also obscure its origin and appearance as it
runs, by altering the HTTP header that contains identifying metadata such as operating system,
browser/version, and referring URL. Kapow Katalyst can use proxy servers, can integrate with third-party
anonymization solutions, and can even detect when access has been blocked during a crawl and
respond by changing proxy servers dynamically during a crawl. Finally, by enabling access to a site's UI,
Kapow Katalyst lets you avoid registering for API access-likely drawing less attention.

Resilience to change
Web sites change daily and some data extraction techniques are highly sensitive to even the smallest
changes in page layout. Kapow's patented Extraction Browser is specifically designed t o be resilient.
The Extraction Browser has deep understanding of Web page structure, and operates on this underlying
structure of a page. Parameter-driven "tag finder" definitions mean that custom tags and syntax errors
do not disrupt processing, and help shield the Extraction Browser from future changes to standards and
technology.

Flexible crawl strategy
Each OSlNT crawler will need to create a crawl strategy tailored to the specifics of the data sources and
mission requirements, and strategies will vary widely. The Kapow Extraction Browser provides the
widest range of options for designing a crawl.
Kapow Katalyst can automatically authenticate to a site, perform complex navigation though menu
structures, supply input for any web form, navigate through multiple pages of output, and scroll through
multi-screen web pages extracting multiple occurrences of data. Crawlers can test for the presence or
absence of data or metadata anywhere on a page, and can change the course of a crawl at any point in
time. Extraction rules can process as many paths for a page as desired.

02011, Kapow Software. All Rights Reserved

In one case Kapow Katalyst was able to penetrate a Flash menu built expressly t o protect a hacker site
from "web scraper" tools, which simply do not offer the same functionality

Multiple language support
Whenever multiple language support is a
requirement, Kapow Katalyst offers built-in
support for single- and multi-byte character
encodings, as well as support for bidirectional
languages such as Arabic and Hebrew. Kapow
Katalyst correctly captures and processes data
from news sites, blogs, RSS feeds, and social
media sites around the world.
The screen image to the right
illustrates Kapow's Design Studio
being used t o extract data from an
Arabic language website. The
Design Studio's graphical editor
allows precise visualization of
selected data in any language and
encoding, enabling point-and-click
extraction of data from any
page-even when the website
does not advertise the encoding.
As a pure Java application, Kapow
Katalyst uses Unicode natively, and data can be saved in UTF-8 or UCS-16. The Kapow Software product
line follows all 118n guidelines, and multi-byte support is fully integrated throughout.

Security
Many websites require a user to authenticate before accessing some or all of the site's contents. The
Extraction Browser has the ability t o enter any user ID and password into the site's authentication page,
just as it can mimic any other action a human user would perform (such as navigating through menus,
scrolling through multiple page outputs, and entering data into forms). These access credentials can be
stored for automated use, or a user can be required t o re-enter them for each crawl, according to
requirements. If a site uses PKI (Public Key Infrastructure), Kapow also has the ability t o use authenticate
using a digital certificate.
Several additional security measures can be applied. You can restrict data transfer to a specific IP
address, port, and socket, and can use SSL for all communication among Kapow Software components.
API access t o Kapow crawlers can support ACL and security, and additional security measures can be
discussed on request. Finally, the Kapow Extraction Browser is stateless, so it can safely reside in your
DMZ: there is no risk of compromising data because no data is retained by the Extraction Browser.
02011, Kapow Software. All Rights Reserved
S

O

F

T

W

A

R

E

Kapow Katalyst is currently in use on several secure networks, including SIPRNET, JWICS, and NMCI.

Automated data de-duplication
Only a small fraction of a typical website's content ever changes once it has been placed online. On a
news site, for example, even though the relative newsworthiness of an article will alter the article's
position on the site over time, the article itself may never be updated-and the Extraction Browser can
detect whether the article has actually changed. Kapow Katalyst can report downstream whenever data
has been added t o a site, changed, or removed from a site, and duplicate data will never be delivered
unless requested.

Offline SnapShot
Studying a private copy of a site-instead of the live site itself-offers

a number of advantages, so

Kapow provides the capability t o take a "SnapShot" of any web page, domain, or list of domains.

A Kapow SnapShot is a complete, fully static copy created by extracting the results of all dynamic page
construction activity. A Kapow SnapShot includes all images, documents, and other linked assets; all
links are automatically rewritten to the local server where the SnapShot has been stored, and all
accompanying metadata is written to XML or into a database for later analysis.
You can access a SnapShot offline without running the risk that the site owner will be alerted t o your
interest due to JavaScript execution of tracking code on the live page. Only Kapow offers both-a static
Snapshot, and the ability to harvest data from a live site.

Advanced Data Normalization
Common on websites is the use of relative dates and time stamps such as "one minute ago" or
"yesterday" and the extraction technology must be able to convert these into usable data. Kapow's
ability t o convert relative times into usable dateltirne stamps is especially important to engage at the
point of capture, because this data cannot be accurately recreated downstream.
Kapow can also access metadata and expose it as data usable by downstream processes. For example,
some sites use color t o indicate the status of an item on the page; other solutions cannot normalize this
type of metadata into data that is useful.

Deep Web Access
Not all data is readily available on standard web pages. Experts estimate that the "Deep Web"consisting of web pages built from queried databases and similar dynamic content-may

be 20x larger

than the "Surface Web" that is accessible by crawling hyperlinks. OSlNT programs cannot afford to
ignore the Deep Web, but harvesting Deep Web data is much more complex than mere link-following.
For example, a given mission may require capturing data from the databases and content management
systems used to support a particular website, and storing it in a local database designed for analyst use.
With i t s ability to navigate complex menu structures and generate input for web-based forms, Kapow
Katalyst can automatically issue multiple queries to display the desired data. Coupled with its ability to
navigate through multiple pages of output and extract data from multiple items on each page, Kapow
02011, Kapow Software. All Rights Reserved

l

Katalyst can readily extract and download entire databases and prepare them for use in any format
necessary. Kapow's de-duplication can be used to automatically keep it current. Other technology,
limited t o following specific links that already exist, will be unable to perform.

Data Integration for OSINT
This section examines requirements generated by the need to integrate raw data into downstream
enrichment and enhancement processes.

Broad-Based Support for Standards
A recent Federal best practice study makes it clear that standards-based products are essential for any
successful OSlNT initiative. Because no single vendor provides everything a mission will require, it will be
necessary for each OSlNT program to integrate the technology it requires-and this integration will be
much easier when proprietary solutions are avoided. Processes that provide raw data t o downstream
enhancement and enrichment technology occupy a special position in the architecture, and bear a
concomitant responsibility to support the widest possible range of industry standards; otherwise,
implementation options will be needlessly restricted.
Kapow supports the widest range of integration standards available. In addition to native support for
XML, HTML, JSON, JMS, and CSV, Kapow can also deliver results into SQL databases and XML
databases-and can create numerous other formats. Data can be exposed as Web services (SOAP or
RESTful) and can be callable via Java or .NET interfaces. Unlike other solutions, Kapow Katalyst-based
crawlers can direct output to any of these formats without being updated-a run-time choice selects the
desired output format.
Finally, Kapow Katalyst can drive content into any Web application via its UI, by automatically
completing the forms used to control the application. No other solution provides this range of design
alternatives for planning your OSlNT architecture.

Near Real-time Access
The same Federal best practice study also points out the desirability of "enriching the dataset as it is
crawled" t o enable near-real time access and improve responsiveness-especially important when
conditions are changing rapidly. This can be accomplished by having the crawler manage some portion
of the enrichment (by invoking whatever enrichment services are needed) before it pushes the enriched
data downstream (either directly into the corpus or via whatever standards-based interface has been
chosen).
Just as Kapow can invoke Web services, Java, or .NET applications during a crawl, it is possible t o
configure Kapow to directly invoke downstream enrichment capabilities during a crawl. This capability
(perhaps most useful during a surgical crawl where a small amount of data needs enhancement) allows
the OSlNT architect to decide based on mission parameters, and not technical limitations, whether realtime enhancement is needed.

02011, Kapow Software. All Rights Reserved
S

O

F

T

W

A

R

E

Partner Ecosystem
Although any successful OSlNT effort will combine a variety of common techniques such as geotagging,
entity extraction, language translation, and other types of enrichment, each OSlNT initiative is unique.
Depending on its mission, each project will be a mix of assembling COTS software and creating missionspecific capabilities, but cost and risk will be reduced by choosing technology that has already been
proven to interoperate.
Kapow Software has an extensive partner ecosystem that includes key vendors and integrators that
specialize in Federal projects, and we have jointly delivered a variety of OSiNT projects. You can feel
comfortable focusing on your OSlNT requirements, knowing that Kapow and i t s partners are dedicated
t o your success.

02011, Kapow Software. All Rights Reserved

Mission Support for OSlNT
This final section discusses issues relevant t o launching and operating your OSlNT capability.

Operational Agility
Most OSlNT missions will need to change over time: sometimes slowly, as mission parameters are
refined through greater understanding, and sometimes rapidly, in response t o conditions on the ground.
Crawlers will need to be created or enhanced as greater understanding is attained, and new types of
data will need t o be extracted as they are found t o be relevant to the mission.
Kapow provides agility in multiple ways. First, unlike other solutions, Kapow does not require a separate
crawler for each site. Sites that share similar characteristics-blogs, for example-can typically use the
same crawler because Kapow operates on the underlying Web page structure, not the presentation
details of a given site. Kapow Katalyst also enables run-time redirection of output t o a different format
(e.g. SQL database, XML, CSV, RSS) allowing a given crawler t o meet the needs of multiple downstream
processes over time. Any crawler can be exposed as a SOAP or RESTful service, or called from .NET or
Java; when combined with Kapow's resilience to web site changes this gives the best kind of agility,
where an external change incurs zero, or near-zero cost.
When it comes time t o extract additional raw data-say, to develop a new surgical crawl-Kapow's
integrated repository makes it easy to share extraction and transformation rules, and the Design Studio
supports rapid, point-and click creation of new crawlers, new extraction rules, and new transformation
rules-often in hours. Analysts can develop their own automated data collection processes, and run
them on demand or on a scheduled basis. Kapow also lends itself t o a Center of Excellence organization
if preferred, allowing extraction skills to be collected in a shared-services model.
Finally, Kapow provides the support you need to handle real-time enhancements t o mission
requirements as conditions change on the ground. For example, in response t o the January 2011
government-initiated Internet blocking in Egypt, Google created www.speak2tweet.com-a service that
recorded voice messages left on a pool of standard phone lines and automatically created a Twitter post
with a link to the recording. Enhancing a Kapow crawler to extract the audio and pass it downstream for
storage and analysis would take under 15 minutes, start to finish.

Scalability
Performance is critical in any large-scale initiative such as OSlNT data capture, and high performance
needs t o be designed in from the start. Kapow Katalyst was designed with linear scalability in mind,
achieved through advanced design and rigorous testing.
Kapow's stateless design improves security, scalability, and availability, and supports the use of internal
cloud-like resources t o enable smooth scaling upward during periods of high load. Each crawl runs
independently, in its own thread, providing the linear scalability needed to support extremely high
volume crawls.

02011, Kapow Software. All Rights Reserved

Production Support
For OSlNT systems that will be run on a production basis it will be necessary to have access t o a high
level of support, both from the vendor as well as from internal IT resources. Kapow provides secure
web-based tools that simplify all phases of IT involvement.
Kapow offers user, role, and group-based control over administration and deployment of the crawler
environment, allowing secure deployment even in a shared-services organization. Fully-automatic
scheduled crawls are provided, along with support for major monitoring standards such as JMX t o
ensure smooth execution and rapid detection of outages should they occur. Kapow Katalyst provides
error logging that can be directed to files, databases, emails, and even custom formats-all completely
secured. Error correction and redeployment are integrated, requiring just a click t o restart.
Finally, as a company Kapow's support team is available 24x7 to assist with software issues.

02011, Kapow Software. All Rights Reserved

Build Your OSINT Capability with Kapow
Kapow's Extraction Browser is a perfect match for the needs of OSINT.
Kapow can fully automate data extraction from any Web site. Without human intervention, Kapow
Katalyst can authenticate, perform complex navigation, carry out sophisticated extraction and
transformation rules, and reliably deliver data wherever it is needed.
Kapow Katalyst is fully standards-based, allowing smooth exchange of data in any required form. Data
can also be exchanged via Web Services, Java, or .NET functions.
Kapow's integrated environment allows you t o build even the most sophisticated crawlers without the
need to develop any code, and test them in real-time-increasing agility and shrinking effort to a
minimum.
An integrated scheduler allows even high-volume crawls to be run on any schedule, and provides
complete production-level support to enhance availability.
Kapow-and our extensive network of technology vendors and integrators-are ready to serve you.
Contact us today, and join the other OSINT projects that are benefiting from Kapow.

Kapow Software
260 Sheridan Ave, Suite 420
Palo Alto, CA 94306
Phone: + l 800 805 0828
Fax: + l 650 330 1062

Email: Sales@KapowSoftware.com
Website: KapowSoftware.com
Blog: KapowSoftware.com/blog
Twitter: @KapowSoftware

107-11 Fleet Street
London, EC4A 2AB
Phone: +44 207 936 9266

02011, Kapow Software. All Rights Reserved

Document Path: ["brochure447.pdf"]

e-Highlighter

Click to send permalink to address bar, or right-click to copy permalink.

Un-highlight all Un-highlight selectionu Highlight selectionh