We are investigation using Playfab for our current game and possible future titles. We currently have AWS servers and are wondering what the expected latency (RTT) between AWS US-west2 and Playfab servers in Azure. Are there any best practices to think about when integrating our AWS Java servers with Playfab?
So, some legacy services in PlayFab sill operate in AWS US-West-2. All the new services we're rolling out operate in Azure (also NW US). So the latency times between those will be fairly minimal. But that said, all the core services are Web API methods - not calls that should be used for latency-sensitive operations. If you have logic that is highly latency sensitive, that should really be done in a custom game server (hosted by us or elsewhere), so that you can a) have those servers running in every region where you need them, and b) have tighter control over the data exchange between server and client - especially if it needs to be at a high rate.
What legacy services are still running in AWS? Just so we know what to expect if we decide to more forward with PlayFab. Also, when making S2S calls, is there any RPS throttling and what is the best practice for a self-hosted game server to communicate with the PlayFab backend when dealing with player data/inventory/etc. We are planning on making the game as server authoritative as we can.
Again, thanks for the quick reply.
Right now, the core API server is hosted in AWS. Since that's where all of your inventory and player data requests will go, that's probably what's going to matter most.
We do apply rate limits to S2S calls as well, but they're about 100 times higher than we apply to clients.
We apply a fairly comprehensive set of server and admin APIs for interacting with player data from a service. You can use any of the APIs listed here except for those under the client section.
2 People are following this question.
