Mavenir launches intelligent video analytics for mobile operators

Nvidia-based AI solution for operators runs on an AI-5G converged platform

Network software creator Mavenir has launched an intelligent video analytics (IVA) that uses artificial intelligence (AI) to video stream analysis to communication service providers (CSPs).

The upshot is that it spots problems quicker, builds up a picture and inspires decisive action, a service CSPs can sell to enterprises. 

The announcement was made to coincide with Nvidia’s GTC conference. 

The system was designed for situations where timely intervention is critical but decision making is complicated by information that’s too fast and too much. The most obvious applications will be in manufacturing, smart cities, entertainment, airports, retail and workplace safety, according to Mavenir. 

Too much information too little time

In smart cities and entertainment AI can deal with large crowds of variables in the shape of cars and people. In airports and in retail it will be tracking the offerings needed to improve customer experiences.

In manufacture and in the work-place it will match regulations to changing situations. The service assesses developments, judges them and makes intelligent interventions.

With camera coverage quickly spreading over every industry, companies are struggling to make sense of all this visual intelligence. This leads to management and storage overload crises. With petabytes of data generated by cameras each day, business users want AI-systems to help their staff spot anomalies, get more details and act quickly.

AI brings clarity on a network slice

Mavenir’s IVA bundles pre-trained deep learning models and pre-configured workflows with AI-infused analytics. This ready-made integration cuts the technology’s payback time from several months to a few weeks, claims the vendor.

With the Mavenir IVA streamlined interface, users are empowered to configure and manage deployments, lowering the costs associated with skilled and scarce data scientists.

The system runs on powerful Nvidia graphic processing units (GPUs) to create instant insights – with waiting time measured in milliseconds. With a cloud-native framework, Mavenir’s IVA is an open system that runs on any cloud infrastructure and, says Mavenir, could be integrated into any CSP’s multi-vendor network.

Mavenir created an opportunity for CSPs

The distributed architecture of IVA uses instant analytics to capture data at edge nodes, and then uploads data content of value to a centralised computing function in the cloud.

These ‘smart edge nodes’ make optimum use of the network’s bandwidth. At the same time the transfer of data to the centralised nodes sets up for cumulative analytics, configuration of alerts and management of the edge nodes.

Mavenir Edge AI, or MavEdge-AI, runs on the Nvidia Metropolis AI-on-5G converged platform, which has an Nvidia A30X converged accelerator and Nvidia ConnectX SmartNIC with 5T-for-5G technology.

First in a series of competitive edges

Nvidia and Mavenir shared a vision, according to Ronnie Vasishta, senior vice president of telecoms at Nvidia. Both wanted to give businesses around the world an intelligent edge and the telecoms software developer made it happen over 5G with artificial intelligence built into edge computing on the cloud.

This is just the first implementation of a series that will appear on Nvidia’s AI-on-5G platform, said Vasishta.