We have an established game, with a backend hosted on GameSparks.

We are considering a port to a new BaaS, as GameSparks has stagnated. There are various options out there and I'd like to hear your opinions / experiences before making a decision.

The game has approx 1000 concurrent users (100k MAU), constantly sending game-state updates to all of their connected friends and to a persisted game state. (The players are effectively connected in a mesh, they aren't partitioned into separate games).

The game is continuous, it never ends.

The features of GameSparks that are key are:

• Persisted gamestate, sharable between players
• Sockets connection for sending/receiving live updates
• Open ended friends lists (not closed player cohorts)
• SDK support for Unity - iOS, Android and WebGL.
• Custom server-side logic

Any advice or insights into alternative backends would be appreciated.

@hplus0603 I feared this answer. We moved from an in-house Java/MongoDB backend on AWS because of the maintenance and running costs a few years ago. I'm reluctant to give up the convenience and flexibility of GameSparks for a ‘harder’ solution again, I like my time off too much!

Thanks for your perspective.

I've come to the same conclusion about other platforms having tried to evaluate likely looking candidates. It seems GameSparks is technically a great fit for us.

I don't have those figures immediately to hand, but the AWS solution was approx 4x the GameSparks cost when we switched over, due to server instances for game and database servers.

I am talking to GameSparks about their future, it may not be quite so bleak after all.

That's a common situation: “The system we are using is terrible, but converting to something else is so much worse.”

There are few off-the-shelf solutions. It is extremely rare that the service given by a company, such as the systems provided by Lumberyard, or what is built into the game engines, is an exact match to your needs.

The big question is the degree of customization you'll need. At a certain point of growth doing it yourself becomes the best option, but it is a long journey to get there. The fact that you currently have 100K MAU means whatever you do you'll be investing a huge effort to keep the service running continuously, so your inertia is an important factor. Often “doing nothing” is the best solution when you can make it work.

That's not quite our situation. Technically the system we are using is fantastic and a great fit for our game. I would praise GameSparks from the rooftops IF they were guaranteed to stick around.

The risk if we stay put is that we lose our backend and have nowhere to go. The risk if we don't is that we spend a lot of development time in the backend and end up with something inferior and more expensive.