Biometric based authentication can provide strong security guarantee about the identity of users. Security of biometric data is particularly important as compromise of the data will be permanent. Cancelable biometrics store a non-invertible transformed version of the biometric data and so if the storage is compromised the biometric data remains safe. Cancelable biometrics also provide a higher level of privacy by allowing many templates for the same biometric data and hence non-linkability of user's data stored in different databases. We define how to measure the success of a particular transformation and matching algorithm for fingerprints. We consider a key-dependent geometric transform that is applied to the features extracted from a fingerprint, to generate a key-dependent cancelable template for the fingerprint. We investigate performance of an authentication system that uses this cancelable fingerprint when a fingerprint matching algorithm is used for detection. We evaluate performance of the system and show the challenges of achieving good performance if the matching algorithm is not modified.