This year, we experienced one of those rare moments that validated every late night, every debugging session, and every coffee-fuelled brainstorming marathon. We hit a 99.97% cache hit ratio—a number that had always felt just out of reach. It wasn't just a technical triumph; it was a clear win for the experience our customers are delivering to their viewers.

The Cache-22 Loop

When content isn't cached locally on edge servers, every request has to travel all the way back to an origin server. This can lead to buffering, delays, and—let's be honest—a poor viewing experience. And when viewers encounter bad performance, they're unlikely to stick around.

Content Delivery Networks start to cache popular content on the edge automatically, BUT without a LOT of viewers, this doesn't happen in every area.

This leads to the cache-22 loop — streaming services need a lot of users to improve their streaming experience, and a bad streaming experience chases away early adopters in any given geography.

For our customers, this wasn't just a technical hiccup—it was a roadblock to growing their audience and platforms.

And for us?

It was personal.

We've all been there, stuck waiting for a video to load. We wanted to break this cycle.

Cracking the Code

This challenge demanded collaboration across all our teams and with our partners at Akamai with months spent diving into logs, spotting patterns, and testing countless ideas.

The solution isn't a single, magical fix — it's the result of a cascade of carefully executed optimisations that include (but are not limited to):

1. Fixing Prefetching
   Prefetching—loading video segments ahead of time—wasn't working as it should. After some deep sleuthing, we realised mismatched segment naming patterns were breaking the process. A few perfectly crafted regex adjustments later (engineers will understand the satisfaction here), and prefetching was back on track. Content started flowing smoothly.

2. Optimising Content Characteristics
   We discovered that our configuration for DASH segments assumed they were much larger than they actually were. This led to unnecessary requests to check object sizes—basically, wasted trips to the origin server. By recalibrating those size parameters, we eliminated the noise and shaved off critical latency.

3. Going Global
   Our customers' audiences span the globe, so their content delivery had to be equally far-reaching. We fine-tuned client characteristics to ensure consistent performance, regardless of location. This adjustment made a world of difference (pun intended).

4. Tuning Segment Durations
   Balancing the fetch timing for current and upcoming video segments might sound simple, but it's a game-changer. By fine-tuning segment durations, we reduced buffering and achieved more stable playback.

The Cache Moment

There's nothing quite like the moment when everything just clicks. The first time we hit that 99.97% cache hit ratio, the excitement in the room was electric. But what truly mattered was the impact this had on our customers.

One of our customers had been dealing with buffering issues in regions far from their origin server. After the optimisations, their user engagement soared. Viewers were finally getting the smooth, reliable streaming experience they deserved, and it was helping the platform attract even more users.

“The relief these optimisations bring is immense, but the opportunities they unlock are even more exciting. Laminar empowers us with technology that not only meets our needs but propels our progress—delivering speed, efficiency, and the momentum to match our ambition,” said Anil S. Singh, Founder & CEO at EvolvX.

That's the part that stayed with us — not the technical milestone itself, but the difference it made for real people.