CPC H04L 63/02 (2013.01) [G06N 5/04 (2013.01); G06N 20/00 (2019.01); H04L 63/102 (2013.01); H04L 63/1425 (2013.01)] | 16 Claims |
1. A method comprising:
receiving, by a server, a plurality of access requests from a requesting entity, wherein the plurality of access requests are received via one or more APIs and are a plurality of requests to access one or more resources in an obfuscated database;
determining, by the server, a requesting entity profile associated with the plurality of access requests;
determining, by the server, a machine learning model corresponding to the requesting entity profile;
generating, by the server, an access sequence, wherein the access sequence comprises an ordered sequence of the plurality of access requests;
determining, by the server, an anomaly score by applying the access sequence as an input to the machine learning model;
comparing, by the server, the anomaly score to a predetermined threshold;
if the anomaly score is greater than the predetermined threshold, preventing, by the server, the requesting entity from accessing the obfuscated database; and
if the anomaly score is less than the predetermined threshold, allowing, by the server, the requesting entity to access the obfuscated database, wherein the obfuscated database comprises a plurality of layers and wherein the one or more resources are stored in a deep layer of the obfuscated database.
|