Thales is using innovative assets from the aerospace industry to boost its Biometric Matching System (BMS) performance while reducing the environmental impact. The BMS is the heart of government digital identification systems. Introducing interoperability with its border management system requires multiple processing of hundreds of millions of biometric database records within 1 to 2 seconds.
To perform data comparison at this scale Gemalto, a Thales company, is applying commercial off-the-shelf Field-Programmable Gate Array (FPGA) technology, originally designed for ultra-low latency applications in high performance computing (HPC) environments in financial and scientific industry, which is compatible with any server and cloud. Specifically, Gemalto uses these FPGA boards for matching hundreds of millions of biometric fingerprint templates which are digital signatures, created from fingerprint images.
This proven solution also allows for much faster data processing and greater matching accuracy, while at the same time limiting infrastructure costs and cutting carbon emissions. Depending on server and system specification, it can require up to 75% less servers and energy overall.
Gemalto’s Automated Fingerprint Identification System (AFIS) and Automated Biometric Identification System (ABIS) are scalable and customizable solutions, providing a range of functionalities for processing, editing, searching, retrieving, and storing fingerprint, palm print, face and iris images and biographic subject records.
With FPGA, ABIS makes it easier for government agencies to run very large and complex multi-biometric solutions with remarkable matching accuracy and speed, enabling States to better protect and serve their citizens. “Our FPGA based solution can cut data centre investment and space overall by more than a half, whilst reducing CO2 emissions by around 50%. Gemalto technology brings new options for governments wishing to prioritise environmental sustainability, without in any way impacting national security.” says Youzec Kurp – SVP Identity & Biometric Solutions at Thales. “The alternative of a pure Central Processing Unit based approach for biometric data matching requires massive computing capacity – even, in some cases, up to four times more servers than Gemalto’s approach – to perform the same transaction.”
Image by Pete Linforth