Asaf Ezra

We actually started completely differently.

We started with load balancing opportunities.

We've seen a huge, huge potential in the fact that load balancers today don't even have any knowledge of what happens on your backend system.

So you can think of gRPC or HTTP load balancing where

Pretty much the only heuristics today is like least connection or least outstanding requests.

But the truth is...

it doesn't mean how loaded the machine actually is, so how much the resources are utilized and for what purpose.

And so we started there.

Unfortunately, it didn't go as planned.

We realized that a lot of people are not very enthusiastic to replace their own very easy to use, infinitely scalable, AWS-backed load balancer.

So we switched to something else that you can start with a single VM.

But the worst case and the cost benefit is like, I'm risking a single VM with this unknown company and I can gain 30, 40, 60% on my performance improvements.

I think our, let's say, median to average customer would be around a few hundreds of machines to a few thousands of machines.

Okay.

That depends on which size the machines are.

That depends if you're seeing the performance improvements that we imagine that you're seeing.

If you're on Kubernetes, you can switch it to the Kubernetes service level or the specific containers that you want Granulate to manage.

So it's not necessarily about how many machines Granulate is running on.

but how many actual course granular is helping you with.

We have a list pricing, but again, I think it depends on the customer itself.