Malicious Website Detection Using Probabilistic Data Structure Bloom Filter

摘要

Bloom Filter is a probabilistic data structure which saves memory space and time efficiently, but the trade-off remains as false positives. It tells us if the value is definitely not in the input stream or maybe in the stream. Since Standard Bloom Filters do not support deleting elements various variants of Bloom Filters have been introduced. Due to the positives of Bloom Filters like compact summarization of streaming data, it has gained importance in applications that use higher volumes of data like in network traffic management, database management and cloud security. In this paper we implement a Bloom Filter to test membership of URLs and provide a warning to malicious websites or access to kid friendly websites. By creating a second Bloom Filter with maximized size we cross verify the query results of the first Bloom Filter to declare with absolute certainty of no false positive result.