Question to other developers

Solar2D Game Networking
#21

@Naomi, thanks much for sharing your experience with a cloud service. I’m guessing this was something like Parse and indeed the scalability offered is very reassuring.

If one chooses to go the other way, say running a NoobHub or Coronium GS instance on DigitalOcean then the cost per API call is fixed but you are pretty much on your own to handle the scalability issues when and if your game goes ballistic. DigitalOcean offers easy upgrades through the server capacity offerings so you can take your $5 server and go to a $10 or $20 or even $40 a month server but at some point you will run out of capacity even in that top tier server.

Then you will need to go to a multi-server setup with load balancing etc and this is where you will find that the cost vs benefits equation would tip over had you been on a cloud service like Parse that does all that for you… Not saying it can’t be done on the DigitalOcean path as there are numerous load balancing and scalability solutions but it just becomes something you need to build and run and it is very time and resource consuming… 

So not knowing your success rate its a good option to go with low fixed cost path on a single server etc but at some point you may need to find an alternative to handle massive success. The way I look at it, if I ever have a game like that then I can afford to bring in some expert paid support and keep the wheel turning. At the end of the day, building and running the backend is not rocket science and there are people you can enlist to get you through the growth pain. The harder thing to do is to make a game that takes you there… At least for me that has been the case so far.

Best of luck! 

#22

Hi Rob,

Yeah I thought it might be too complicated. As for Autolan, I have tried to contact the developers through this forum and they have ignored my PM. Another person and I have a thread going where they found it was an issue with the Sockets: http://forums.coronalabs.com/topic/48123-autolan-local-multiplayer-issue/ We have put in a bug report, but got no response.

As for charging for limited time use multiplayer. I wonder if in the end that might hurt the app as a whole. I feel today’s users are quite entitled when it comes to spending money on apps. I think it might be hard to sell a limited time use IAP for anything in a game. In the end, it really comes down to how popular your app becomes.

I was looking at CoroniumGS as well and was very interested. But for those of us that would have to still pay for a server on AWS or w/e does it still amount to the same issue of a snowballing monthly cost? I’m feeling there is no easy solution to this as a single/small indie developer.

#23

[quote name=“AppOwlDev” post=“264686” timestamp=“1410544868”]I feel today’s users are quite entitled when it comes to spending money on apps. I think it might be hard to sell a limited time use IAP for anything in a game.[/quote] So true. I think the approach to sell this might be to give a free install a set period of free multiplayer. Say, one week free after which you turn off multiplayer and offer IAP to restore functionality. A free trial period per se.

#24

I struggled for months about what do for a multiplayer solution because I wanted to do cross-platform multiplayer. In the end, I decided to avoid the 3rd party multiplayer solutions that would allow me to do cross-platform because of the cost. I’m currently developing a solution that uses GameCenter on iOS devices and Google Play Game Services on Android devices, though this isn’t easy since Corona uses turn-based on GC and real-time on GPGS.

This problem would be moot if we could get Corona Labs to implement GPGS for iOS, and also support turn-based on GPGS. Earlier this year, Google announced that GPGS is now available on iOS, which will allow for free cross-platform multiplayer play. Check out this thread for more info:

http://forums.coronalabs.com/topic/45644-google-play-games-gets-cross-platform-multiplayer-for-ios-and-android/

I would encourage you to throw some votes behind this issue in Corona Labs’ feedback system. Below is the applicable feedback case:

Turn-based Multiplayer and iOS support for Google Play Game Services

 

If you have the votes available, I would also encourage you take a look at these two:

 

Real-Time Multiplayer support over Game Center on iOS

 

Support all GameCenter features

#25

Hey, @drewns, am I right to understand that GPGS does offer turn based multiplayer.  Not just real time?  I guess by allowing Google Play version to access Apple Game Center??  (BTW, I’m not personally interested in realtime multiplayer on iOS…)

Naomi

#26

Yes, GPGS has both turn-based and real-time multiplayer now, and so does GameCenter. Like you I’m not as concerned with real-time multiplayer (I’m using only turn-based multiplayer for now). However, it would be nice if Corona gave us access to both turn-based and real-time on both systems.

The biggest benefit to Corona developers though would be for Corona Labs to add the GPGS iOS SDK to the gameNetwork library. Basically, this would allow someone on an iOS device to login to GPGS (using their Google+ account–just like on an Android device), which would allow for cross-platform multiplayer play. It would allow Corona’s GPGS gameNetwork functions to be run on an iOS device.

It may also be possible to let the user login to both GC and GPGS to record achievements and leaderboard scores, and choose which game network to use for multiplayer, though I can’t say for sure if that would work.

#27

That’s huge.  I cast my vote.

Naomi

#28

I agree! Thank you for voting!

#29

Voted but not holding my breath! With all they have on their plate I doubt this will be a priority for Corona Labs anytime soon.

Thankfully, Coronium Cloud does all I need in for turn based and now with Coronium GS for realtime… Wohoo. Much thanks Develephant.

#30

@ksan, I’ve been reading/hearing truly great things about Coronium.  It’s just that I’d like to minimize the use of third party solutions wherever & whenever I can.

Naomi

#31

Can’t find fault in that strategy. Having less moving parts generally make the machine simpler and more reliable. With all the high level abandoned 3rd party products of late one has to be more and more cautious… 

However, I think Coronium is here to stay and grow. I am betting all my future projects on Coronium. You can’t go wrong with Develephant’s amazing work. 

#32

Coronium looks great, and will definitely fit the bill for many developers. And I certainly respect Develephant’s work. At the same time, it would be nice if we had the GPGS option as well.

I’m hoping that since the GPGS feature request is almost Top 10 now (at #11 as of this moment) Corona Labs might take a hard look at it. Seems like a very do-able feature for them to add, since they already have some GPGS features implemented.

#33

My plan is to offer local multiplayer at launch.  If I get a high enough adoption and interest, I will look into online multiplayer.  

This is for exactly the reasons you cite above.  I didn’t want to ship the game with online multiplayer and then have to cancel it due to cost.  It seems like the players would hate me for that or at least quit playing.  I felt like multiplayer was really important to the success of my game so I made this design decision so that once I launch, the player will have access to what they originally downloaded indefinitely at no additional cost to me.

Since I will have local multiplayer, it won’t be a big leap to go to online multiplayer.  At least, not nearly the leap to go from no multiplayer to online multiplayer.  My data is already used to being bundled, shipped around and unbundled.  My UI accommodates other players info.  So far it seems this is the lion’s share of the work.

Here is another thing that I’m interested in, I’m hoping that local multiplayer will drive word of mouth.  The player recruits his/her friends to play because they want to play local multiplayer.  I’m still working on it so I can’t verify any of this, but I thought it was worth mentioning.

I really wish there was a good LUA api for Bluetooth.  It seems that a lot of public networks don’t support users connecting on the LAN.  I want to experiment with using one of the devices as a hotspot as well as the server.  I may be able to get that to work, we’ll see.

#34

Hey Mark, can I ask what you are using to incorporate local multiplayer? I’ve been looking all ovr but there doesn’t seem to be any solutions other than the broken Autolan.

#35

I just wrote it myself using lua socket.  I figured it out using these references:

This one is a competitors website, I hope I don’t tick anyone off for listing it but it is a good explanation of how to use UDP and TCP.  I went from not knowing anything about it at all to being able to code my own module reading this.  It says Gideros but it is just LUA code:

http://appcodingeasy.com/Gideros-Mobile/Using-LuaSocket-in-Gideros

Here is more info.  The explanation is not as good but it’s not bad code:

https://github.com/Overtorment/NoobHub/blob/master/client/lua-corona/noobhub.lua

Here’s something very similar to the link above, possibly taken from it:

http://stackoverflow.com/questions/19831520/lua-socket-connection-for-corona-sdk

Here’s the official documentation:

http://w3.impa.br/~diego/software/luasocket/

The gist of it is that you use UDP to let the clients find the server and then you connect them with TCP. I use a button that says “find servers” and “find client” that runs a loop for a few seconds so they can find each other. Then there is a handshake process to make sure you are actually talking to a client of the same game and then they just start sending messages back and forth using TCP. If you read that first article, what I just said will make sense.

My actual game is quite large and I didn’t want to break it while I hacked away at network communications so I just made an app that could link up two devices and then I could send text messages back and forth. Once I had it working I integrated it into the game.  The core of network communications is enabling a computer to send strings to the other computer which can then be interpreted.

It took me roughly eighty hours to get it working outside my game, including time to find resources and study TCP and UDP. I say that so you don’t just throw up your hands and go WTF when you read the links. It might take you less time, I don’t have a background in computer science but I’ve been working in LUA for a while. You kind of have to get your head wrapped around it so… be patient. Once you understand the terms, it’s not that tricky, it’s just a self contained kind of thing with it’s own jargon. Once you have written some acceptable code for your game you can easily use it in a future project which is a consolation.

Maybe I can write something up for Corona so they can have their own local networking tutorial.  We’ll see if I can find some time.

#36

We welcome guest tutorials!

Rob

#37

Hi Mark,

Thank you so much for the info! I will start looking into it right away. I would LOVE to see a tutorial too!

#38

I removed my custom multiplayer feature from my word games precisely because of the reason you noted – i.e., too expensive.  I supported it for a few months, then removed it entirely from my games (until the Game Center & GPGS came along to fill the gap.)

Unless the app/game monetizes extremely well with tons of users who are willing to pay hundreds of dollars, chances are, it will cost you bundle to keep it going.  I think forking out hundreds of dollars every month as if giving gifts to some random strangers is a bit too much for indie developers.

So, when the Apple Game Center and Google Play Game Services enabled us to offer multiplayer feature, I was so happy to add the multiplayer feature back in.  (Even though they are not the same, it works, and the best thing is, we don’t have to pay for the use of their servers.)

Naomi

#39

Hi Naomi, thank you for confirming what I thought might be a lost cause for me. I will probably just concentrate on using the native utilities (Game Centre and Google Play). 

#40

Yeah, the thing is, unless you know with the absolute certainty that the app/game will NOT have lots of users, you’d have to be prepared for massive usage, and that sort of preparedness costs $$$.  

I just don’t think anyone would be okay with not being prepared – because not being prepared is the same as willing to let the servers to fail when the app/game succeeds.

So, when unsure, it’s better to test the multiplayer feature of the app/game on Game Center and Google Play.  You can always switch to your custom solution when the app/game proves to be super duper successful.

Naomi