We have been working on a game for the past year on the GameSparks platform.
I recently started to review the Game Data Service documentation and decided to try to test adopting it in my game.
It is 5-8x slower in execution time (shown in the Profiler in Test Harness) than the exact same operations in Mongo. I'm not talking about anything fancy here - I'm talking about just creating an empty collection, retrieving 1 row, updating 1 row type operations. I haven't even try testing more robust/complex operations yet.
I'm wondering if anyone else has found this issue, and how they resolved it. Did you wind up just not adopting GDS? Did you adopt anyway? I have reached out to their support, who basically have told me this is just the way it is.
To me, a 5-20x increase in execution time is completely unacceptable. I have had to reconsidering using GameSparks entirely, although that seems like a dramatic option. I am really hoping they get their act together. Support hasn't really said much in response to this, other than that they are making architecture changes which includes possibly adding new AWS features.
I was even considering using DynamoDB over HTTP as my persistent storage, mainly just due to fear of them dropping support of Mongo updates and support, although I may just stick with runtime collections for now as I'm grandfathered in.
Thank you for any feedback.