SciTransfer
FAST-STREAM · Project

AI-Powered Video Streaming Optimization to Eliminate Buffering and Lag in the Last Mile

digitalPilotedTRL 7

Imagine a highway where traffic jams happen right at your exit, making your delivery late. This technology acts like a smart GPS that predicts these jams every few milliseconds and instantly finds a faster way to send data. It ensures your video calls and movies stay crystal clear without those annoying pauses, regardless of your Wi-Fi or 5G quality.

By the numbers
10s of milliseconds
granularity of rate selection optimization
The business problem

What needed solving

The 'last mile' of internet delivery is volatile, causing video buffering and lag. This leads to customer churn and lost revenue for streaming and communication providers.

The solution

What was built

CompiraEdge software for HTTP/TCP and QUIC/UDP delivery, integrated with a CompiraCloud data platform and SaaS dashboard.

Audience

Who needs this

OTT Video Streaming Services5G and Fiber Network OperatorsVideo Conferencing PlatformsContent Delivery Networks (CDNs)
Business applications

Who can put this to work

Telecommunications
enterprise
Target: Internet Service Providers (ISPs)

If you are an ISP dealing with customer churn due to poor video quality on mobile and Wi-Fi networks — this project developed CompiraEdge software that optimizes data delivery to reduce re-buffering and resolution drops.

Media & Entertainment
enterprise
Target: OTT Streaming Platforms

If you are a streaming service dealing with users leaving because of long lags during live sports or events — this project developed a machine learning-based congestion control system that maximizes the quality of experience.

Enterprise Software
any
Target: Video Conferencing Providers

If you are a communication tool provider dealing with unstable connections for remote workers — this project developed a solution that adjusts transmission rates every 10s of milliseconds to keep calls smooth.

Frequently asked

Quick answers

What is the cost or pricing model for this technology?

Based on available project data, the specific pricing is not disclosed, but the solution is delivered via a SaaS dashboard and CompiraCloud data platform.

Has this been tested at an industrial scale?

Yes, the project conducted large-scale field trials and pilots with global partners including Telefonica, Lumen, Conversant Solutions, and Medianova across Spain, Chile, Peru, USA, Mexico, Vietnam, Malaysia, Europe, and the Middle East.

Who owns the IP and how is it licensed?

The technology is based on Performance-oriented Congestion Control (PCC) developed by Compira co-founder Prof. Michael Schapira and Prof. Brighten Godfrey; licensing details are not specified in the report.

How difficult is it to integrate into existing networks?

Integration is simple because the solution requires no changes to the existing network or applications to function.

What is the timeline for deployment?

The project ran from 2022-05-01 to 2024-09-30, focusing on scaling the technical solution for a commercially viable product.

Consortium

Who built it

The project is led by a single SME, Compira Labs Ltd, which indicates a highly focused commercial drive. While the consortium consists of only 1 partner, the business risk is mitigated by the involvement of high-profile industrial pilot partners like Telefonica and Lumen, showing strong market pull.

How to reach the team

Contact Compira Labs Ltd in Israel

Next steps

Talk to the team behind this work.

Contact us to explore licensing opportunities for PCC-based congestion control.