Is it feasible to track all visitors to the United States and then build a Data-driven DSS?
In response to the September 11, 2001 World Trade Center and Pentagon attack, the U.S. Congress required Attorney General John Ashcroft and the U.S. Immigration and Naturalization Service (INS) to develop a plan for a National Security Entry-Exit Registration System (NSEERS). The first phase of NSEERS was implemented by the U.S. INS at selected ports of entry throughout the
When fully implemented NSEERS will be operational at more than 100 airports and 200 land and sea ports of entry. In a
· Fingerprinting and photographing at the
· Periodic registration of aliens who stay in the United States 30 days or more.
· Exit controls that will help the Immigration and Naturalization Service to remove those aliens who overstay their visas.
Ashcroft asserted on
NSEERS is a massive information systems project. For both political and technology reasons it seems that the system may not capture the same data for all visitors.
In assessing the feasibility of the system, I developed a list of questions:
1. What decision support is needed?
2. What data should be stored by NSEERS?
3. How large would a comprehensive database need to be to track all visitors to the
4. What architecture would one create for real-time data entry and access to information in such a database?
I'll briefly share my answers to these questions, feel free to suggest additions and to challenge the assumptions I have made.
First, decision support is needed at the border to determine if a person should be admitted to the
Decision support capabilities will need to be developed. More decision support is need than lists of people who are unlawfully in the
Second, what data should be stored? Clearly some personal data -- name, address, citizenship, date of birth, height, weight -- needs to be stored; a photo and a thumb print. Then a "transaction log" of entries and exits needs to be maintained. For many visitors they will have only two entries in the log in a calendar year - an entry and an exit. Some frequent visitors at the Canadian or Mexican borders might enter and exit 5 times a week.
Third, the database would be very large. The transaction processing database will need to be loaded regularly (probably daily) into a data warehouse for analysis and decision support. According to FBI materials, it takes about 45 KB to store a single fingerprint. A photo of visitors would take about 10 KB. Personal information would take about 5 KB. For 35 million visitors, each at 60 KB of storage, that would equal 2.1 Terabytes per year. A data warehouse would probably need to store 5-10 years of data.
Fourth, the architecture will require communication with a centralized database for fingerprint comparisons, but it is possible to store some data at each port of entry and upload it in batches. Data entry hardware for NSEERS needs to be in place and network connections are needed at each port of entry. Multiple input devices will be needed at many ports of entry. Maintenance of the hardware at ports of entry can probably be outsourced. The software and hardware will need to scale to manage a very large database. The TPS database will grow rapidly and the query demands could become very demanding on the system once it has million visitor records.
Is it feasible to track all visitors? YES. Will NSEERS be an easy TPS/DSS project to implement? NO. Implementation and OPERATION of the system must be managed with care and sensitivity. Visitors to the
Teradata staff, "Harrah's Entertainment, Inc.," NCR-Teradata, 2000, at URL http://teradata.com.
U.S. Department Of Justice Press Release, "Foreign Visitors to
The above is from Power, D., Is it feasible to track all visitors to the
Last update: 2005-08-07 11:26
Author: Daniel Power
You cannot comment on this entry