17LIVE keeps live entertainment running at the speed of its audience

Since launching more than a decade ago, 17LIVE has grown into one of the region's largest live-streaming platforms, with a business built almost entirely around what happens in real time. As that audience grew across APAC, so did the pressure on 17LIVE's infrastructure. Every gift sent, every leaderboard update, and every piece of content served needed to happen at a much larger scale, without slowing down. 17LIVE turned to Redis on Google Cloud to meet that bar and to give its engineering team room to keep building.

Challenge: Real-time demand was outpacing 17LIVE's database

Live entertainment runs on moments, and moments do not wait for slow technology. On 17LIVE, gifting drives the platform's economy, leaderboards drive its competition, and both spike hard during a live show. Add search and content matching to the mix, and the platform's infrastructure was handling several real-time workloads at once, all under the same tight latency expectations.

17LIVE's original architecture processed these interactions directly against its database. That worked at a smaller scale, but as the audience grew across the region, it started to show cracks. Gifting transactions lagged behind the action on screen during high-traffic events. Leaderboards fell out of sync during competitions, undercutting the sense of real-time competition that keeps viewers engaged. Search and content matching, which viewers expected to feel instant, slowed at exactly the moments that mattered most.

Eric Hsu, SVP of Engineering at 17LIVE, led the effort to fix this. His team needed an approach that could absorb sudden spikes in gifting and engagement without adding latency, and that could scale with 17LIVE's growth across APAC rather than becoming a recurring bottleneck.

Solution: One real-time layer for gifting, leaderboards, & personalization

17LIVE moved its most latency-sensitive workloads onto Redis Cloud, replacing direct database calls with a real-time layer built for its busiest moments. Gifting transactions, leaderboard updates, and session data now run through Redis, keeping response times under one second across all of 17LIVE's critical paths, even during the largest live events. Running Redis on Google Cloud let 17LIVE's team deploy quickly within its existing infrastructure, without standing up and managing a separate caching layer on its own.

To support personalization and content discovery, 17LIVE also uses Redis for vector search. Rather than relying on simple keyword matching, the platform uses vector search and processing pipelines to match viewers with relevant content and to help surface the right creators to the right audiences in real time. Hsu adds, "By using Redis, 17LIVE implements near real-time interactions in its live streams."

Underlying all of this was a simpler goal: give Hsu's engineers a foundation that could absorb sudden spikes without manual intervention, so the team could spend its time building new features instead of managing infrastructure during peak traffic.

Impact: Faster experiences, happier viewers, & more than a decade of reliable uptime

The shift paid off immediately. 17LIVE's critical paths, including gifting, leaderboards, and search, now respond in under one second, even at peak traffic, and database load dropped by 50 percent. That headroom gives 17LIVE's engineering team more room to build without worrying about the next traffic spike.

Viewers noticed the difference. User complaints fell by 30 percent, follow rates on creators rose by 10 percent, and event spending, the core of 17LIVE's gifting economy, increased by 5 percent, each pointing to a more engaged, more satisfied audience.

That combination of speed and reliability has held up over more than a decade of live, always-on operation without a major service disruption, on a relationship with Redis that stretches back to 2016. For a platform built entirely around real-time interaction, that track record matters to 17LIVE's business as much as any single metric: it means the team can keep growing its audience without rethinking its architecture at every stage. "Redis supports our critical business path to create high scalability for our diverse situations," Hsu said.

“

"Powered by Redis, 17LIVE delivers near real-time services to create a seamless and immersive experience for both Livers and viewers."

Eric HsuSVP of Engineering, 17LIVE