Enhancing securing architecture of signalling system No.7 with emphasis on machine learning techniques to detect attacks

Abstract

Signalling System No. 7 (SS7) is used in GSM/ UMTS telecommunication technologies for signalling and management of communication. It was designed on the concept of private boundary walled technology having mutual trust between few national/ multinational operators with no inherent security controls in 1970s. Deregulation, expansion and merger of telecommunication technology with data networks have beaten the concept of boundary walls hence increasing service providers, entry points and interfaces to SS7 network making it vulnerable to serious attacks. SS7 exploits can be used by attackers to intercept messages, track a subscriber’s location, tape/ redirect calls, adversely affect disaster relief operations, drain funds of individuals from banks in combination with other methods and send billions of spam messages. This thesis provides a comprehensive review of SS7 attacks with detailed methods to execute attacks, methods to enter SS7 core network and recommends safeguards against SS7 attacks. It provides implementation of machine learning concepts Vs rule based filtering to detect anomalies in SS7 network, a template for rule based filtering of specific SS7 messages and a conceptual defence model for the defence of the network.