Realistic Expectations?

narfi · March 2014

Been playing around with Photon for a while now as a hobbyist trying to teach myself.
I have a game scene with ~75 animals/monsters wondering around or on patrol paths each with a PhotonView.

I have no problem connecting and running around in the scene with one character, but if a buddy joins from across the country (I am in alaska, he is on the east coast) then it gets super laggy and eventually kicks one of us out. Once the other disconnects, then the remaining person has a smooth game again.

Is this too much for the cloud to handle, or am I doing something wrong either in terms of the networking or game coding?
Idealy I would like to be able to have 5, 10, 20 players able to play together in a scene like this, so if I need to be teaching myself a new approach would appreciate any direction in this

I have tried researching your MMODemo that comes with your server, but there seems to be little to no documentation about it and perhaps is even an abandoned project?

I have used your self hosted server some, but just the loadbalancing (myCloud) so that I could work without a connection, when testing the game with friends I reset it to the cloud just so I don't have to worry about routing and all of that.

The transition to using the self hosted server was intimidating till I realized that there is no difference between the loadbalancing and the cloud except for the address in the settings. How steep a learning curve would/will it be to use the mmodemo or start putting together my own?

Appreciate any ideas, thoughts and input you might have

thanks,

narfi

edit: I have seen mention of area of interest or a limit to just network updating things close enough to the player to matter, is this possible in the cloud or just with the server? And if possible with the cloud, where is this refrenced so I can learn about it?

narfi · April 2014

Bumping this up for some direction.

Borzi · April 2014

No, that does not seem unrealistic. If the lag comes from movements, then I would highly recommend quill18's tutorial on PhotonNetworking. Very useful to solving lag issues (although you might have already adapted this approach).

https://www.youtube.com/watch?v=oZ7TQjWhtEw

Hope this helped!

Edit: If Tobias says an MMO doesn't work, then it probably doesn't....but i think that the tutorial might still help you with lag issues

Tobias · April 2014

The Photon Cloud is not built for MMO style of games and if you have more than 16 or 32 characters to sync, the room based logic of the Photon Cloud is not suitable.
I'm not sure if you only have so many NPCs or if you want more than 50 players in the scene, too.
For the NPCs and monsters, you could minimize the synchronization by sending only their next waypoint. Or the next few waypoints with a time when they should arrive there. Then each client can move the monsters around without actually syncing their current position 10 times a second.

The MMO Application from the Server SDK is not compatible with PUN and kind of abandoned.
Afaik, it should still work but it's not very well documented and much more complex due to areas, interest, etc. It's not a very streamlined product but the basis for a proper seamless-world type of game, which is (very) complex to build.

narfi · April 2014

Tobias wrote:

if you have more than 16 or 32 characters to sync, the room based logic of the Photon Cloud is not suitable.

That is unfortunate, but what I had feared. While not necessary right away, I would like to work in the direction of 15 to 20+ players in a single scene.

Tobias wrote:

The MMO Application from the Server SDK is not compatible with PUN and kind of abandoned.

Ok, that is what it appeared like from researching the forums here.

I think you are missing a big opportunity for hobbyists like myself in this aspect.
While I am realistic enough to understand that a true MMO type of game is beyond my knowledge/skill base, there is no motivation to learn going forward if the bar is too high.

Your site discusses concepts such as areas of interest and how important they are, and yet you have no compatibility between PUN which I had seen as the 'stepping stone' and the more complex aspects of your server solutions.

I will continue researching to see what other options there are, or If I need to change my goals, or what my best approaches are.

Thanks for taking your time to answer my questions

narfi

Tobias · April 2014

Don't get me wrong: I do appreciate your input. I just don't really agree.
We might be missing an opportunity but once you dive deeper into this topic, you will notice that it's by no means easy to do. And as if effort and complexity wouldn't be enough, there is a high probability that any implemented solution will not cover some customer's needs and we begin from scratch.

If you feel like it's a missed opportunity, then I can only encourage you to implement a workflow for this and sell it in the Asset Store!
You can do this on top of Photon, if you wanted to. It's certainly up for the task, if you come up with the fitting logic.

At the moment, we concentrate on turnbased games which can be played asynchronously over a longer time.
We might revisit the topic of interest management later on.

Sorry we couldn't cover your use case.

Realistic Expectations?

Comments