Optimizing Media Quality with the Network for Real-time Communications through Artificial Intelligence

By Allie Mellen on May 21, 2018
read

Using artificial intelligence in conjunction with the network to optimize media quality is a more abstract idea than many direct applications of AI. Nevertheless, it has the potential to seamlessly improve real-time communications. Predicting network resources, minimizing round-trip time, and actively monitoring the network are three tasks humans are currently responsible for. However, artificial intelligence could take over these tasks to advance real-time communication and enhance call quality.

Predicting Network Resources

In modern-day networks, bandwidth is determined via capacity planning and manual, human adjustments. This leads to eventual human error or poor planning. Instead, artificial intelligence could be used to predict network performance and needs based on previous trends. For example, the network will recognize when a video call is about to begin and provide sufficient resources to bring the optimal media quality.

Improving Network Routing

High round-trip time (RTT) during a video conference can have an adverse effect on media quality. In order to minimize the RTT per video conference while maintaining quality, it is critical to use the best media server possible for each specific call. This is dependent on a number of factors, including the location of the participants in the call, the location of the server, and the server status. If the media server is located far away from the end-users, it introduces a higher RTT value and is known as hairpinning.

In order to minimize the RTT, artificial intelligence can be used to choose what media server is best used for a specific call. It can use predictive forecasting to anticipate media server status and enhance call quality while maintaining a low RTT.

Monitoring and Measuring the Network

Monitoring the network is a frustrating and tedious task. It can take network domain experts days to inspect all of the data related to the network. In contrast, artificial intelligence can be used to understand and interpret terabytes of data in minutes. Being able to monitor the network quickly and effectively enables teams to quickly understand and fix what specifically is causing performance issues.

Similarly, monitoring, debugging, and resolving issues with real-time communications can be a daunting task. There is a lot that can go wrong, and it can be hard to predict and identify the source of an issue. Instead of pure human intervention, artificial intelligence could be used to debug and identify issues in a similar way.

Improving Quality of Experience

Assessing Quality of Experience (QoE) within a call is difficult, since a lot of the factors rely on the degree of delight or annoyance of the user of an application or service. In order to address this, at callstats.io we address three key points: network settings before the call, during initialization of the call, and during the call. Addressing and assessing network characteristics, especially before the call and during initial setup, can help understand how the QoE will be affected and make changes to improve it.

Optimize, our AI-powered product, is built for just that. Optimize uses artificial intelligence algorithms to provide media and network settings that deliver optimal audio and video quality for each device, connection, media, and network setup. We leverage data from millions of conferences to bring optimal audio and video quality to every interaction.


Tags: Real-time Communications, Artificial Intelligence, Networking