Dispute Analysis and Reporting in VoIP Wholesale CDR using Big Data Techniques

Abstract

The VoIP industry has grown immensely and it is predicted to grow with double the rate in coming years. The impact of VoIP growth on economy is treasured. The growth of VoIP industry has also increased the volume of data, which is challenge for processing. VoIP security vulnerabilities and lack of equipment cause billing disputes and fraudulent attacks, thus affecting VoIP wholesaler’s profits. To reduce the economic losses, these challenges need to be addressed in an intelligent and efficient way. This study proposes an intelligent and adaptive rule based reconciliation process to resolve real time billing dispute with minimum loss. Disputed real time CDRs are analysed to generate adaptive rules for catering dynamic data. These rules are used to categorise the CDR into 5 categories. A summarized report is generated at end of analysis which assist the negotiation process to resolve billing dispute. The complexity and volume of data affects the execution time of the reconciliation process. Spark, a distributed processing framework, is used to reduce the execution time. The distributed processing reduces the execution time 11 time than a non-distributed solution. The performance of proposed solution is evaluated against solution of local software house and the proposed solution has detected 38% more billing disputes in less time as compared to the existing solution.