An Algorithm-Independent Cluster Computing Platform for Highly Scalable Biometric Matching
Astra™ is a cluster computing platform used for large-scale biometric identification and deduplication using fingerprint, face, and iris recognition. It is a highly scalable platform that performs one-to-many search or one-to-one match against large stores of biometrics and other identity data. It does so by deploying biometric data and matching algorithms across a cluster of multiple computing nodes.
Extremely large biometric databases (tens of millions of records) cannot be stored or efficiently searched on a single computer. Distributing the data and biometric comparison tasks across multiple machines enables even extremely large databases to be searched in only seconds. Astra enables not only massive biometric search tasks but complex filter, search, match, and link operations critical to data preparation and quality assurance functions such as identity resolution and data deduplication.
Astra makes use of an open source distributed computing framework that is in use in a diverse variety of large-scale systems, and thus is field-proven, reliable, and well-supported. Astra is highly fault-tolerant with integrated failover capability, and includes a browser-based system monitoring dashboard that allows administrators to know the performance of the system at all times and be alerted when problems arise.
Astra configures a multi-node computing topology and deploys the algorithms and data for fingerprint, face, and iris recognition to each node, then manages the execution of the millions of operations and results in such a way that maximizes the utilization of the machines and processors. The platform is independent of the algorithms in use, so long as access to 1:1 matching capability of the algorithm is provided. The platform is able to run one or more biometric and text matching algorithms in either staged or parallel search configurations.
Astra Core provides system configuration, management, and monitoring. System topology and workflows are configured. The biometric data cache is managed. Biometric matching algorithms are configured, managed and deployed. A browser-based system monitoring dashboard provides alerts and statistical information about the operation of the system. The cluster controller executes commands and computations across multiple computing nodes. It deploys biometric and text matching algorithms and data to each node.