question

I believe I'm just confusing you further, as those were references to two completely different ways to have server-side logic. What I would recommend is that you read up on the docs for those services on their sites, as that would be the best way to learn about them. Realistically, I should just keep my comments down to the PlayFab services. And along those lines, I can't speak to the costs on any service outside PlayFab.

If you need a higher rate of checks than we discussed above, and you need server authority to do that, the recommendation would be to use custom game servers. Relay services do not provide you with server authority to check for and prevent bad data (cheating).

I'm not sure what "original custom server" refers to. You can upload any .Net compatible server build in a zip to our hosting and we take care of the scaling in the regions you activate. The pricing is the AWS EC2 hosting cost, plus 10% to cover our costs, if you go with our legacy hosting. We're finalizing the costs for the V2 service now, but it will be competitive with the current model, at least - very likely lower.

The new service is Azure-based, and so provides for more regions. It also can host Linux servers, and can run in a Container model. And yes, it does have a matchmaker - our new matchmaker is a more sophisticated queued matching system, and is also currently in preview.

No, if you use hosted game servers, PUN would be superfluous.

And no, our pricing starts at $0.008 USD per MAU, and we discount that price at higher MAU levels - only the first 250K are at that price. Talk to our sales team (sales@playfab.com) if you want to talk through pricing for high MAU games.

Ultimately, the success of your game is up to you. We provide a wide range of tools needed to run games, including all the live game operations components needed to drive events, re-engagement campaigns, customized behaviors based on user segmentation, A/B testing, and more.

But running a game is like any other business. If you were to open a shoe store and people came to look at your shoes but never purchased any, that wouldn't have any impact on whether you owe the landlord of the mall for the space, electricity usage, etc. We provide a free tier so that developers who are trying to figure out how to make games that will be successful and convert players can launch games that don't use high levels of resources (which have real world cost, even at lower levels). Where titles need more resources, we do have to pass those costs on, as financially we don't have any other choice.

If you feel strongly that you don't yet have a feel for whether your title will convert players, I would strongly advise you to scale back your feature design to fit within the limitations of the free tier, without requiring the kind of high frequency, real-time data that you're describing.