Project Sentry is a counterterrorism data-mining application used in Afghanistan, this is an ongoing series on the program. This post is just about the software and basic identification of the users, next post we'll get into the kill list.
Things added to things, as statistics, civil history, are inventories. Things used as language are inexhaustibly attractive.
- Ralph Waldo Emerson, Representative Men
In a hub-and-spoke data architecture, a central entity collects information from the outer spoke elements, processes and synthesizes the information into an intelligence product, and loads it into a central repository for use by all elements within the system.
The basic SENTRY federation concept was to "set up server at each participant’s site and link these servers together in a private cloud. Each participant's "data will remain in their portion of the cloud and we will bring experts in to analyze the information (in bulk) and develop tailored rules and algorithms to clean, format and standardize each participant’s information. ...As new data is loaded into the participant’s area of the cloud, automated programs will measure new data as it arrives to look for changes and identify new relationships."
More images on the design of SENTRY
- Ralph Waldo Emerson, Representative Men
The Sentry Federation
The SENTRY project was developed to suit the data ingestion needs of a number of US and international agencies. The main US entities are the DoD, DEA, and Treasury Department, and included are international members of the ISAF in Afghanistan, as well as the Israeli Defence Force -- all stakeholders in the Afghanistan Terror Finance Cell, ATFC.
The variety of data sources as well as diverse needs of the ATFC consumers presented a special problem
The cell analysts, located in Kabul and Baghram had to "manually manipulate the raw intelligence data that is collected in support of law enforcement actions, military operations and targeted financial measures executed in coordination with the Afghan counterparts," according to notes taken by the civilian lead for the project.
That raw data is "country specific financial and micro level information that requires unique processing and transformation" in order for the analysts to identify suspicious income and revenue streams and turn the data into actionable intelligence.
That raw data is "country specific financial and micro level information that requires unique processing and transformation" in order for the analysts to identify suspicious income and revenue streams and turn the data into actionable intelligence.
Further complicating the analytical effort," wrote the project lead, "is the informal (Hawala) money handling information which requires scanning, Optical Character Recognition (OCR), and translation."
The solution, SENTRY, a "federated system where participants can maintain control of their data, yet data quality is raised to levels equivalent to spoke and hub systems."
In a hub-and-spoke data architecture, a central entity collects information from the outer spoke elements, processes and synthesizes the information into an intelligence product, and loads it into a central repository for use by all elements within the system.
The basic SENTRY federation concept was to "set up server at each participant’s site and link these servers together in a private cloud. Each participant's "data will remain in their portion of the cloud and we will bring experts in to analyze the information (in bulk) and develop tailored rules and algorithms to clean, format and standardize each participant’s information. ...As new data is loaded into the participant’s area of the cloud, automated programs will measure new data as it arrives to look for changes and identify new relationships."
"The initial Sentry proto-type consists of a central repository to store CTFI information, data binning, OCR and machine translation capabilities, data cleansing and standardization algorithms, and analyst tools and utilities managed by workflow processes."
More images on the design of SENTRY
Comments
Post a Comment