A comparison of centralized versus distributed architectures in biometric access control systems

The need to control access to physical facilities or computer databases is increasing each year. Identity verification is implicit in this control, i.e. one must be able to confirm the identity of the individual at the door before access can be granted. In the past, automated identity verification was difficult. Early access control systems either relied on manual means of identity verification (e.g. a security guard recognizing each face), or they simplified the problem to verifying the identity of some physical object (e.g. an access card) rather than the individual. The first technique was expensive. The second carried with it the assumptions that this object was always attached to its owner. This was a poor assumption, and the loss or sharing of access cards represents the biggest weakness in access control systems today. In the last decade, the technology has become available to allow true verification of individual identities. This technology is based on a field called "biometrics", in which distinctive biological features are measured for each individual. These features can later be checked against the individual requesting access, and that individual's identity can be accurately confirmed. This paper will introduce and compare biometric access control systems based on centralized or distributed architectures. It will compare these two architectural approaches, outline their relative strengths and weaknesses, and discuss applications where one or the other is more appropriate. The need to control access to physical facilities or computer databases is increasing each year. Identity verification is implicit in this control, i.e. one must be able to confirm the identity of the individual at the door before access can be granted. In the past, automated identity verification was difficult. Early access control systems either relied on manual means of identity verification (e.g. a security guard recognizing each face), or they simplified the problem to verifying the identity of some physical object (e.g. an access card) rather than the individual. The first technique was expensive. The second carried with it the assump tions that this object was always attached to its owner. This was a poor assumption, and the loss or sharing of access cards represents the biggest weakness in access control systems today. In the last decade, the technology has become available to allow true verification of individual identities. This technology is based on a field called "biometrics", in which distinctive biological features are measured for each individual. These features can later be checked against the individual requesting access, and that individual's identity can be accurately confirmed. This paper will introduce and compare biometric access control systems based on centralized or distributed architectures. It will compare these two architectural approaches, outline their relative strengths and weaknesses, and discuss applications where one or the other is more appropriate.