Spam Over Internet Telephony Detection Research Articles

Efficient Detection of Spam Over Internet Telephony by Machine Learning Algorithms

Recent trends show a growing interest in VoIP services and indicate that guaranteeing security in VoIP services and preventing hacker communities from attacking telecommunication solutions is a challenging task. Spam over Internet Telephony (SPIT) is a type of attack which is a significant detriment to the user’s experience. A number of techniques have been produced to detect SPIT calls. We reviewed these techniques and have proposed a new approach for quick, efficient and highly accurate detection of SPIT calls using neural networks and novel call parameters. The performance of this system was compared to other state-of-art machine learning algorithms on a real-world dataset, which has been published online and is publicly available. The results of the study demonstrated that new parameters may help improve the effectiveness and accuracy of applied machine learning algorithms. The study explored the entire process of designing a SPIT detection algorithm, including data collection and processing, defining suitable parameters, and final evaluation of machine learning models.

Systems and methods for SPIT detection in VoIP: Survey and future directions

In recent years, VoIP (Voice over IP) telephony has shown a tremendous increase in the number of subscribers due to today's affordable telephony rates and flexible use of Internet technology for voice communication. At the same time, a proportional increase was exhibited in VoIP spamming and SPam over Internet Telephony (SPIT), which are forms of abuse and frauds that can have severe consequences and financial losses for both the service providers and subscribers. This paper surveys, reviews, and discusses the state-of-the-art detection and mitigation techniques and systems for VoIP spamming and SPIT. The paper highlights reasons and motivation behind such abuse and fraud, and it discusses the primary challenges in devising an effective and efficient anti-SPIT detection solutions. Moreover, the paper outlines shortcomings and limitations of existing solutions, and it identifies future research directions to aid in further improving and enhancing effectively and efficiently the detection and mitigation of SPIT and spamming.

Rapid detection of spammers through collaborative information sharing across multiple service providers

Spammers and telemarketers target a very large number of recipients usually dispersed across many Service Providers (SPs). Collaboration and Information sharing between SPs would increase the detection accuracy but detection effectiveness depends on the amount of information shared between SPs. Having service provider’s exchange call detail records would arguably attain the best detection accuracy but would require significant network resources. Moreover, SPs are likely to feel uncomfortable in sharing their call records because call records contain user’s private information as well as operational details of their networks. The challenge towards the design of collaborative Spam over Internet Telephony (SPIT) detection system is two-fold: it should attain high detection accuracy with a small false positive, and should fully protect the privacy of users and their service providers. In this paper, we propose a COllaborative Spit Detection System (COSDS)—a collaborative SPIT detection system for the Voice over IP (VoIP) network where service providers collaborate for the effective and early detection of SPIT callers without raising privacy concerns. To this extent, COSDS relies on a trusted Centralized Repository (CR) and exchange of non-sensitive reputation scores. The CR computes global reputation of users by aggregating the reputation scores provided by the respective collaborating SPs. The data exchanged to the CR is not sensitive regarding users privacy, and cannot be used to infer the relationship network of users. We evaluate the performance of our system using synthetic data that we have generated by simulating the realistic social behavior of spammers and non-spammers in a network. The results show that the COSDS approach has better detection accuracy as compared to the traditional stand-alone detection systems. For instances, in a setup where spammers are making calls to recipients of many SPs, COSDS successfully identifies spammers with the True Positive (TP) rate of around 80% and false positive (FP) rate of around 2% on a first day, which further increases to 100% TP rate and zero FP rate in three days. COSDS approach is fast, requires a small communication overhead, ensures privacy of users and collaborating SP, and requires only few iterations for the reputation convergence within the SP.

Two-stage SPIT detection scheme with betweenness centrality and social trust

Detecting SPIT (Spam over Internet Telephony) is an urgent demand with growing voice communication services. Chaisamran et al. proposed a trust-based SPIT detection scheme and it is superior to other trust-based schemes since it is valid for unknown users. However, this scheme might misdetect a call from low-frequent users as SPIT gradually in time, since they seldom call and thus their social trusts get decreased with time. In this paper, we propose a two-stage SPIT detection scheme using BC (Betweenness Centrality) and social trust to decrease misdetection of a call from low-frequent users as SPIT. BC indicates user's centrality in the entire network and the value of BC against legitimate users gradually increases with time even if users seldom call, whereas it does not increase against spammers since the connection between a legitimate caller and a spammer is hardly established. As a first stage, we use BC as a feature to correctly identify a call request from a low-frequent user. Then we judge whether a call is from a legitimate caller or a spammer by using social trust. By the computer simulation, we show that our scheme improves the false positive rate while maintaining high true positive rate.

Using data mining approaches to identify voice over IP spam

SUMMARYWith its rapid growth in popularity and attendant personal and commercial use of voice over Internet protocol (VoIP), spam over VoIP has emerged as a significant problem. In this paper, we first applied genetic algorithms to weight predetermined features of VoIP calls. Based on the weighting technique, we then proposed a mechanism for the online prevention of VoIP spam by the k‐nearest neighbor classification. Differing from the traditional k‐nearest neighbor algorithm, this paper used genetic algorithms to weight features to reflect the different degree of importance of each feature. VoIP is a real‐time communication service, as compared to off‐line e‐mail, and VoIP spam is also known as Spam over Internet telephony (SPIT). The system proposed in this paper has been named PSPIT that denotes Prevention of SPIT. Once spam is identified, the proposed PSPIT immediately updates the built‐in blacklist on the Session Initiation Protocol (SIP) server to block that specific user from sending SPIT again. The n‐fold cross‐validation was applied to evaluate PSPIT for unknown SPIT detection; in the case of n = 2, the best overall accuracy was 94.7%, whereas in the case of n = 4, the best overall accuracy was 93.1%. Copyright © 2013 John Wiley & Sons, Ltd.

Securing SIP-based VoIP infrastructure against flooding attacks and Spam Over IP Telephony

Security of session initiation protocol (SIP) servers is a serious concern of Voice over Internet (VoIP) vendors. The important contribution of our paper is an accurate and real-time attack classification system that detects: (1) application layer SIP flood attacks that result in denial of service (DoS) and distributed DoS attacks, and (2) Spam over Internet Telephony (SPIT). The major advantage of our framework over existing schemes is that it performs packet-based analysis using a set of spatial and temporal features. As a result, we do not need to transform network packet streams into traffic flows and thus save significant processing and memory overheads associated with the flow-based analysis. We evaluate our framework on a real-world SIP traffic—collected from the SIP server of a VoIP vendor—by injecting a number of application layer anomalies in it. The results of our experiments show that our proposed framework achieves significantly greater detection accuracy compared with existing state-of-the-art flooding and SPIT detection schemes.

OntoSPIT: SPIT management through ontologies

VoIP enables new ways for communication. At the same time, it provides new means, in terms of transmitting bulk unsolicited messages and calls, namely SPam over Internet telephony (SPIT). In this paper, we propose a conceptual model, based on an underlying ontology, which describes the SPIT domain. The ontology provides capabilities, such as modeling the SPIT phenomenon in a SIP-based VoIP environment, a common understanding of SPIT domain, as well as reusable SPIT-related knowledge interoperability, aggregation and reasoning. We demonstrate that the proposed ontology, combined with a set of SPIT identification criteria, as its underlying axioms and rules, could enhance the correlation and management of SPIT incidents. It could also support SPIT detection, thus facilitating the better protection of VoIP environments in a holistic, cooperative, and effective way.