Spam and unsolicited communications have been bothering users of emails since years already.
In the recent years spam evolved and started being delivered by means of voice over phones, both
legacy and IP-based. While systems for email Spam filtering are already quite effective (if
configured correctly), protection systems against Spam Over Internet Telephony (SPIT) and real-
time data in general are still in the development phase. One big challenge in this area is not only
the real-time nature of the data to be analyzed, but also the need for per user configuration of the
protection system.

Stage at NEC Laboratories Europe (Heidelberg, Germany)

Download proposal

VoIP Security

The most important factor inhibiting VoIP implementation has changed from a lack of budget to concerns about security. Security attacks in the VoIP environment are an increasing concern as VoIP developments are growing dynamically around the globe. Most systems have security holes that if left unmonitored risk to compromise the success of VoIP itself towards the replacement of classical telephony. Currently deployed Intrusion Detection/Prevention Systems are not able to counter VoIP attacks (VoIP SPAM, DoS, Call re-routing, Call hijacking, etc.). Main objective is to study and design a system able to counter VoIP SPAM and possible VoIP attacks. Such a thesis will study and develop auto-learning pattern based systems in the context of preventing VoIP systems to be broken by intruders and misbehaving clients. Validation of results is expected to be done either on simulation or on real implementation (or a combination of both).

Pre-requisites:

• TCP/IP knowledge: mandatory
• VoIP protocols/architecture basic knowledge -- basically SIP and H.323 (they could be acquired during the thesis work).