Big data analysis on secure VoIP services

VoIP users increase each day. However, the documentation on the behavior of VoIP applications is still lacking. The needs to understand and generalize the behavior of applications like Skype, GoogleTalk and SIP based applications grow each day. There are many factors that influenced the performance of a VoIP application such as bandwidth, packet loss rate, delay, jitter, codec type and CPU power of the end devices. The user experience of the service is important since, VoIP is a real time application running over the best effort internet. Since VoIP data co-exist with other data on the internet, extracting, transforming, loading and analyzing the selected VoIP application is a challenge. We design an instrument to do data collections, data massaging, data analysis and data interpretation of large amounts of network packet. The result shows that GoogleTalk, Skype and Express Talk are more sensitive to the impairments due to packet loss rate and jitter rather than to the impairment due to delay. Bandwidth and other resources like a de-jitter buffer and a gateway's CPU and memory are important in order to produce a good quality VoIP service. The lack of these resources would result in several packets lost before they reach a destination or the packets arrive too late to join the other packets in the de-jitter buffer at the destination gateway. The gateway would drop these packets if the de-jitter buffer is full or not enough memory or CPU powers to process the packets. A gateway closer to the receiver end decapsulates IPSec or TLS packets. The gateway also decodes the voice packets before the packets entering the receiving machine. High throughputs do not imply high Perceptual Evaluation of Speech Quality for Wideband (PESQ-WB) scores. The throughput size is determined by the codec type and the security features that are implemented on the infrastructure.