I'm neck-deep in evaluating ZPA against competitors like Palo Alto Prisma Access and have hit a wall with their sales team on a basic question. Their pricing is per "application," but their definition seems to shift depending on who I'm talking to.
From what I've gathered so far, an "application" in ZPA is essentially a unique FQDN or IP/port combination that you're making privately accessible. But the devil is in the details.
* If I have a single web server hosting multiple internal sites (say, `hrportal.corp.com` and `finance.corp.com`), is that one application or two? The rep hinted it could be two if they are distinct DNS names.
* What about a single server with multiple ports? Is `appserver.corp.com:443` and `appserver.corp.com:8443` considered one or two apps?
* Does the backend server type matter? Is a legacy on-prem Oracle database on a single port priced the same as a modern SaaS-like microservice?
I need the real-world, contractual definition, not the demo slide. Anyone who has been through a ZPA procurement cycle:
* What exactly were you charged for per seat?
* Were there any surprises when you onboarded a new service?
* Is there a hard limit in the contract, or is it purely usage-based?
Their competitor's pricing is per seat, flat. I'm trying to understand if ZPA's model is genuinely more granular and fair, or just a way to obscure the true cost until you're locked in. The negotiation is stuck on this point.
Don't pay list price
Yeah, I'm curious about this too, especially the part about different ports on the same server. I was looking at a trial and the configuration step seemed to treat each unique "domain:port" as a separate segment. So `appserver:443` and `appserver:8443` would likely be two.
Does the pricing tier maybe change that definition? Like a more expensive plan might group them?