GameLift architecture

Guys, good afternoon! I wanted to know a little more about all this architecture to close some ideas and I have some (a little silly) questions that I want to ask.
When you create a fleet of instances in Amazon GameLift, what does the game session represent?
Is it an instance of the instances given by the fleet?
And the player session? Is it another extension but given by the game session? I’m right?

Hey there @LuisZ , hopefully this explanation clears up some of your questions regarding terminology/architecture:

  • A GameLift fleet is a collection of managed Ec2 instances.

  • Each instance runs a number of serverProcesses (up to max 50). You specify the number/configuration for the serverProcesses by passing in a RuntimeConfiguration to the CreateFleet API.

  • The number of active, healthy server processes is the # of available gameSessions that you can run on your fleet.

  • When you create a gameSession (either via CreateGameSession / StartGameSessionPlacement / StartMatchmaking APIs), you reserve one of those healthy server Processes for usage

  • Once you have an ACTIVE GameSession PlayerSessions can be created on it. Each PlayerSession has an PlayerSessionId, and provides connection information for a player to connect to the ec2 instance your GameSession is running on.

  • The # of PlayerSessions that can be created on a GameSession is specified on the API call used to Create the Session:

1 Like