Executive Summary

Executive Summary

Biometrics is the science of establishing the identity of a person based on or his or her physical, chemical, or behavioral characteristics. It is a rapidly growing field with many applications.  Some examples include verifying the identity of a person attempting to access a computer network, or someone conducting a transaction with an Automated Teller Machine (ATM). Furthermore, in modern society, the constant threat of terrorist attacks underscores the need for a reliable large scale biometric identity system capable of accommodating a large number of individuals.

The limitations associated with today’s large scale biometric systems are that they are generally inflexible and not optimized for use within an enterprise.  Many biometric systems are procured based on their image capture and match algorithm capabilities, with little thought given as to how the system will fit into an organization’s existing system architecture, or how the biometric information will be used within a particular agency’s business process/structure.  As a result, many biometric systems are developed in a stovepipe fashion with little or no interoperability with other biometric systems. Furthermore, proprietary vendor algorithms provide limited system flexibility.

Team Biometric Enterprise Architecture (BM EA) seeks to investigate the limitations associated with current biometric enterprise architecture implementations, and ultimately provide some alternative implementations that will generate improvements in system flexibility, interoperability, and performance.

Team BM EA followed a structured system engineering approach to developing and evaluating alternative architecture implementations.  The team documented an “As-Is” biometric architecture implementation along with requirements for an alternative “To-Be” biometric architecture implementation. Team BM EA used a systems engineering modeling tool to capture the functions and data flows for the “To-Be” architecture implementation. The “To-Be” implementation will eliminate stovepipe, redundant components in line with our stakeholder requirements.  Communication will reside on a common, standards-based, data portal, with more efficient and interoperable communication between elements within the architecture. 

Through research into methods for accomplishing the stakeholders’ key goals of flexibility, interoperability, and open architecture, the team selected service-oriented communication architecture back by cloud computing “commodity” hardware.  These technologies deployed with a flexible architecture to take full advantage of these technologies were shown through analysis and simulations to meet the key stakeholder goals as an effective cost point.  The “To-Be” will allow for the flexible use of multiple vendor match algorithms, prioritization of transactions through the system, virtualization of servers to provide flexible hardware processing resources to meet immediate transaction needs, and the agile allocation and de-allocation of processing power to cost effectively meet surge needs during peak periods or during high threat conditions. This new architecture provides the advantage of better overall performance over a wider range of different transaction types and scenarios particularly under ever shifting workload, mission priorities, and budgets.

Team BMEA developed a performance model to compare the performance of two different types of implementations through a hypothetical border crossing application and to demonstrate how an engineer can take this new architecture and develop a flexible system optimized to a particular application.

Based on initial results obtained from the performance model, Team BM-EA recommends that agencies/organizations attempting to introduce biometric enterprise architecture within their business construct implement Service Oriented Architecture (SOA) like technologies on cloud computing “commodity” hardware to improve system interoperability, flexibility, and performance at an improved price point. The team also explored the use of an agent-based model.

Furthermore, results from Team BM-EA’s cost modeling indicate that significant low-risk savings can be realized by switching from the As-Is implementation.