Jeff Baxter

Or your optimal point is X. Instead, it's more, I'm an A300 with this many workloads on.

They're ramping at this rate.

This is what I'm predictably performing at.

And you're sort of self-diagnosing yourself, right?

It's kind of like,

I'm running this marathon.

This is how my body is feeling.

This is the pace I've been at.

So based upon prior experience and knowing what my body is capable of, I'll be able to finish this marathon in X minutes or I have this much capacity left in me, right?

It's not hard-coded.

It's dynamically generated.

On the QoS minimum, the one thing I will mention that's kind of cool is I did mention it's built into the entire system.

And basically what we do is prioritization.

We know as every workload enters the system, if it has a QoS minimum,

And what the priority is, and we essentially set a kind of deadline to that individual IO or that individual message.

And all the way down from the protocol layer through the waffle layer down to the individual disc layer, we're aware where that deadline is when we need to get that IO back out and we prioritize accordingly.

So the analogy I've been using is, you know, basically the security line at the airport.

And they don't do as much anymore.

And I think maybe it's because of preferred traveler lines and TSA pre and all this other good stuff, right?

But