By Calvin Biesecker
The Defense Department’s next-generation biometrics repository that went online in late January is already exceeding the capabilities of the system it replaced and is also showing itself to be a reliable platform that is already accepting routine fixes and upgrades, DoD and Army officials said yesterday.
The Next Generation Automated Biometric Identification System (NGA) is making about 10 percent more matches than the prototype ABIS and 96 percent “lights out” matches versus 88 percent for ABIS, Greg Fritz, product director for DoD Biometrics within the Army’s Program Executive Office for Enterprise Information Systems, said at a briefing. Lights out matches refer to automatically and definitely either matching a biometric, such as a fingerprint, submitted by warfighters against the database or not producing a match because the repository doesn’t have a matching biometric, he said.
Increasing those “lights out” matches means there are fewer submissions that can’t be resolved automatically and therefore require an evaluation by an expert examiner. So far the number of transactions requiring human intervention have decreased by more than 50 percent through the use of NGA in the past month, according to an information paper released by the DoD Biometrics Program Management Office.
That’s part of the time savings that NGA provides to the warfighters who want to know as quickly as possible the identity of someone that may be applying for access to a United States military facility in Iraq or Afghanistan.
NGA moves the response time to “minutes and seconds” from the “hours and minutes” that ABIS provided, Col. Ted Jennings, project manager for DoD Biometrics, said at the briefing.
Version 1.0 of NGA, as the current iteration of the system is referred to, is providing customer response times that are 14 to 28 times faster than ABIS, according to the information paper released by Jennings’s office.
The NGA is the “Defense Department’s central authoritative enterprise repository where we process and store, share, match and mange biometrics data,” Fritz said. He also said that it is “secure, stable and scalable.”
There have been some unscheduled outages of NGA that Fritz attributed in one case to a power outage and another to an issue with the palm print algorithm. But those were minor and last week during routine maintenance of the system fixes and security patches were installed that demonstrated NGA could be improved on a “regular, recurring basis,” he said. Monthly releases of fixes and upgrades are planned, he added.
The NGA was officially turned on Jan. 29, about 30 hours prior to the planned go-live date (Defense Daily, Feb. 13). While ABIS was a repository for storing and matching fingerprints, NGA is multi-modal, meaning it also will store, match and manage iris and facial images and palm prints. The system also features more advanced matching algorithms and multi-modal fusion technology that can automatically produce definitive matches from two or more biometrics, such as fingerprint and iris, when either one of those by itself would have an inconclusive result, Fritz said.
Fritz said he is unaware of any biometric database in the world that incorporates the multi-modal fusion search and match capability.
The advanced matching algorithms have already been proven, Fritz said. Use of NGA recently helped lead to the capture of an enemy insurgent whose biometrics had previously been taken or captured 11 different times, whether through latent prints taken from exploded bombs or attempts to access U.S. installations, but had been able to evade ABIS. NGA, on the other hand, pulled those separate occurrences together and linked them to a single person, he said.
One Army official said that the multi-modal fusion technology was proven in trial periods during development of NGA. However, Fritz pointed out that the storage of biometrics other than fingerprints is brand new to the database so it will take time to build these records within the repository and generate matches.
ABIS had grown to about 3,000 fingerprint records but NGA is designed to handle over 4 million records. The system features blade server technology supplied by IBM [IBM], one of the subcontractors to NGA prime Northrop Grumman [NOC], that enables the growth in the database as well as the future incorporation of new technologies and biometric modalities.
Fritz praised Northrop Grumman’s efforts on NGA so far, saying the company has done a “great job” as the systems integrator and has “beat the bushes in industry to find the best technology.” In addition to IBM, other key subcontractors include L-1 Identity Solutions [ID], who is supplying the multi-modal fusion capability, Oracle [ORCL], Symantec [SYMC], Ideal Innovations, and others.