Page 288 - Automated Fingerprint Identification Systems (AFIS)

APPENDIX B: INTERNATIONAL ASSOCIATION FOR IDENTIFICATION 273

3. TEST APPROACH

3.1. STANDARDS-BASED

To communicate across jurisdictional and vendor boundaries, standards are
essential. In developing our tests, we adhered to the ANSI-NIST Data Format for
the Interchange of Fingerprint Information. We used the FBI Electronic Fingerprint
Transmission Specification as a standard, but found it necessary to make a few
modifications based on the specific needs of cross-jurisdictional use. These are
outlined in Appendix C. And lastly, we used the FBI’s CJIS Wide Area Network
Interface Specification to specify the mode of transmission, specifically, the use of
SMTP with MIME partitioning.

3.2. INTERNET TESTING

Last year, ComnetiX, a software integrator who participated in our testing, sent
a suite of test messages to vendors via the Internet using SMTP with MIME par-
titioning, and vendors sent test messages back. ComnetiX confirmed the
vendors were WSQ and ANSI-NIST compliant by nature of the fact they were
able to decipher the messages. Higgins & Associates, International, then con-
firmed the messages were EFTS and ANSI-NIST compliant with help from FBI
and NIST personnel. This year, we repeated the Internet testing, adding the
latent transactions.

3.3. NLETS TESTING

NLETS is the common name referring to the National Law Enforcement
Telecommunications System message switching system created in 1968 for and
dedicated to the criminal justice community. NLETS includes a wide area frame
relay network (installed in 1997). For the IAI testing, we were concerned only
with the frame relay network, not the message-switching computer.
Two of the sites (NC and AZ) connected to the NLETS frame relay network
using existing circuitry to access their state’s NLETS network at a speed of 56
KBS. The Cogent site in Ontario, CA and Aware in Bedford, MA used a dial-
up line running at 14.4 KBS. The dial-up connections required modems and
routers in Ontario and Bedford in order to connect to the NLETS Phoenix
location.
While 14.4 KBS certainly sufficed for the testing where we compressed latent
images using WSQ compression, this speed is rather slow for sending uncom-
pressed images, as is desirable for the transmission of latent prints.