The client is a well known Silicon Valley network monitoring company that was selected by a leading European Telecom to deliver an effective converged network measurement system.
As a leading European telecom was rolling out new VoIP, video and data services, it found itself responding to customer complaints about quality and performance. Numerous traditional network monitoring solutions in place were neither detecting the underlying causes, nor tracing them back to individual endpoints and network segments. In order to continue rolling out these innovative new services, the telecom needed a more effective converged network measurement system.
There are several challenges in measuring VoIP and video quality. Typically, the quality measurements need to come directly from the end points. When there are millions of endpoints in use, it is not practical to collect, store and analyze all the information produced by the monitoring devices. Statistical sampling or using synthetic sessions may lead to false negatives and give the provider an impression of normal performance. Having customer complaints as the first signal of poor performance was unacceptable.
Furthermore, traditional network monitoring solutions are not able to identify IP flow-based issues that cause poor quality and performance. Even when traditional monitoring could indicate network problems, there was no easy way to match up which VoIP calls and video streams were being impacted. Furthermore, IP metrics alone are not a proper quality indicator for video performance. An understanding of whether progressive or adaptive streaming is being utilized is critical to understanding the quality of experience (QoE).
Meeting the Challenge
DataArt was engaged to develop a highly scalable solution, capable of handling millions of sessions per day, that would be able to monitor and compute R-Factor (objective call quality) and QoE from centralized monitoring points. By developing signaling decoders for SIP, SCCP, H323, H225 RAS, Megaco, MGCP, UNIStim and several video codecs, DataArt was able to reconstruct and match individual quality metrics and provide near real time visibility for the telecom to proactively manage issues. For video services, DataArt developed new methods to determine the type of video streaming in use and relate the IP metrics to generate a realistic QoE result.
In order to store and process all of the information, DataArt developed an innovative method of storing data in the product databases. This solution effectively reduced write operations by a factor of 10X and decreased data retrieval time by 3X.
DataArt’s converged network visibility and monitoring solution enabled the telecom to:
- Ensure quality of service and continue the rollout of their innovative offering
- Increase customer acquisitions and retention
- Promptly detect underperformance and avoid negative impact on the business
- Develop procedures to ensure the future health of the converged network
- Comply with regulation relating to call quality
- Decrease infrastructure costs by allowing the telecom to properly provision their network
FreeBSD, Linux, MySQL, C++, Perl, Commercial VoIP monitoring components.