Case studies / Satt Live

Live streaming with sub-second latency, and hundreds of thousands of bids at once.

Satt Live is a platform where products are presented on a live stream and sold by auction. We built its streaming infrastructure and real-time bidding system.

Client
Satt Live
Sector
Live commerce and auctions
Scope
Live streaming infrastructure, real-time auction system
Engagement
Custom software development
Satt Live
Under 1 s
stream latency
1M+
viewers the stream can scale to
100K+
people can bid at the same time
CLIENT
In a live auction, the moment a viewer sees on screen and the bid they place have to line up. If the stream runs a few seconds behind, viewers bid on an auction that has already closed; if bids arriving in the same second get out of order, nobody accepts the result.

Common live streaming platforms split the video into segments to reach large audiences, and with this method the stream often reaches viewers more than ten seconds late. That delay goes unnoticed in a concert stream, but it breaks fairness in an auction where seconds matter. Solutions that cut latency, on the other hand, struggle as the audience grows. For Satt Live we needed both at once.

On the streaming side, we built a distribution layer that keeps latency low and scales horizontally as viewers grow; new capacity comes online automatically as load increases. We also designed the infrastructure to keep operating costs low.

On the auction side, we built a real-time bidding system that puts bids arriving in the same second into a single, correct order. Every bid is validated on the server, and the highest bid and remaining time are sent to all viewers at the same moment. When two bids arrive at once, there is no dispute over which came first.

Goals and outcomes

What did we achieve with Satt Live?

Goal
Outcome
Bring live stream latency below one second.
Bids line up with the moment viewers see on screen; nobody bids on a closed auction.
Build a streaming infrastructure that grows with the audience.
The stream can scale to millions of viewers, and latency doesn't rise with audience size.
Keep the bid flow smooth even at the busiest moment.
Hundreds of thousands of people can bid at once; every bid is validated and placed in a single queue.
Run this infrastructure at low cost.
Capacity adjusts to the number of viewers, so the infrastructure runs at low cost for its streaming volume.