Abstract:
Methods and systems for authenticating a user device employ a database of global network latencies categorized and searchable by location and calendar date-time of day usage, providing network latency by geography and by time. The database is constructed using voluminous daily data collected from a world-wide clientele of users who sign in to a particular website. Accuracy of the latency data and clock skew machine identification is made practical and useful for authentications using a service provider-proprietary, stable reference clock, such as an atomic clock, so that internal clock jitter of a service provider performing authentications does not affect the network latency time and clock skew identification of user devices. Increased authentication confidence results from using the database for correcting network latency times and user device signatures generated from the clock skew identifications and for cross checking the authentication using comparisons of initial registration to current sign in data.
Abstract:
Methods and systems for authenticating a user device employ a database of global network latencies categorized and searchable by location and calendar date-time of day usage, providing network latency by geography and by time. The database is constructed using voluminous daily data collected from a world-wide clientele of users who sign in to a particular website. Accuracy of the latency data and clock skew machine identification is made practical and useful for authentications using a service provider-proprietary, stable reference clock, such as an atomic clock, so that internal clock jitter of a service provider performing authentications does not affect the network latency time and clock skew identification of user devices. Increased authentication confidence results from using the database for correcting network latency times and user device signatures generated from the clock skew identifications and for cross checking the authentication using comparisons of initial registration to current sign in data.