doubts with unet pricing

Tbh I tl;dr half this conversation.

I was wondering with all this talk about ccu costs, if running a direct ip connection setup without the matchmaking service will cost me money? I would love to learn online multiplayer, but I'll never actually use it if it costs me $. I was always under the impression that it didn't but I guess I should make damn sure.

 

If you setup a direct ip connection if the IP is not the final server (the server isn't under NAT) it won't connect because you don't have a way to traverse NAT

 

So basically players will fail connection, even with the correct IP inserted if I don't have Nat punch-through? I only tried direct ip with two computers in the same house thus far. I guess players could use hamachi, but that's super inconvenient. Darn.

 

Yes, I suppose that hamachi will be an option but if you're planning to release a game with that.. I don't recommend it too much XD

I'm looking at this asset for the punchthrough https://www.assetstore.unity3d.com/en/#!/content/58948 it works for PC, mac and linux but you have to have an operational server (AWS of amazon for example) or If you don't mind to pay 80$ one time and you're planning to release on Steam I suggest Photon Bolt https://www.photonengine.com/en/Bolt It seems great also.

 

So I read up on Bolt and it seems to me that it includes the same Structure as Unet. (Local and Remote Clients), but also provides additional components such as automated Animation syncing, etc. This sounds really convenient and worth investing time in, but then comes my question on Zeus. Zeus seems to be the answer to my Nat-punchthrough and matchmaking issues. But to what cost? Is there a price model outside of the $80 asset? (I'm assuming Zeus and Bolt are somehow related to the same product).

 

@Azmar If you have to "support a lot of people" and you can't afford the costs, then you don't have a valid monetization strategy to go with your game design. Your game design has to take into account the costs involved in making and supporting the game. If you can't afford to pay somebody to host your servers, and you don't have the in-house (or personal) resources to run a server yourself, then you need to reconsider the kind of game you are developing, or how you intend to make money from that game. Network game design is fraught with difficult technical problems and with infrastructure problems and with cash flow problems. They are literally the most difficult kind of game you can make, IMO, and require the most diverse set of skills. I hope that you are able to plan a path to success!

 

Zeus is a factilitator to the NAT Punchtrough but if you use the Steam servers (you must be greenlighted before to have an appID) you don't need Zeus, they're planning an out of the box integration with Steam servers.

 

Looking at the posts here it seems like the biggest issues ppl are having is that they really need to look on what data they send for syncing. I cant understand how ppl need even 1k of data/second for just updating the player.
-Do you really need to sync the whole transform?
-Do you really need to sync all values in a vector?
-Do you really need the precision of a double?

In my game (which is an RPG, so its a bit friendlier on bandwidth, i know) im down to just 60 bytes/s per player/mob.
-Only syncing X and Z, which i compress into ints for 1dm precition, Y is handled by the collider stuff
-Rotation is only Y, compressed down to a short, for 0.1 degrees precition
-Actual movement is synced by syncing animation changes, movement are synced just 10 times/ sec to conserve bandwidth.

Just some stuff to keep in mind : )