You've identified the core trade-off correctly, but you're underestimating the financial opacity of Cato's "managed outcome."
>How deterministic is Cato's steering during backbone congestion events?
It's deterministic for them because they control all variables. You're buying an SLA, not visibility. The consistency is achieved through engineered over-provisioning, a cost you're paying for but can't audit. Ask them for 95th percentile utilization metrics during your PoC. If they refuse, you have your answer.
On Versa's operational overhead, it shifts from managing policies to managing a performance model. You don't just set a policy for VoIP and forget it. You'll spend cycles tuning health probes and thresholds so the Director doesn't make technically correct but user-experience-poor choices. It's a dedicated role.
APAC PoP comparisons are almost irrelevant for Cato; you're buying their map. With Versa, you're aligning with your existing cloud vendor commitments, which can be a cost and compliance advantage.
Your cloud bill is 30% too high
Your read on Cato is interesting. That "trust their backbone" part is what I'm trying to wrap my head around.
>How deterministic is Cato's steering during backbone congestion events?
This seems like the key question, right? If you can't see their telemetry, you're trusting their definition of "congestion." Maybe it's fine, but you're locked into their problem-solving.
For Versa's operational overhead, I'm curious if anyone's quantified it. Is the overhead in the initial policy setup, or is it the constant tuning of those application health checks to stop it from making weird choices? Seems like you'd need a dedicated person for that model tuning, which is a hidden cost too.
Containers are magic, but I want to know how the magic works.