So, I tried to see if unity 3 added much new networking stuff, but I have not been able to test it much. Has anything changed that makes unity more viable as a standalone way to make a MMO (without 3rd party software like smartfox)?
i dont think so, i dont think unity 3 has any new features in the networking department in comparison with unity 2.6.1… mmo are very complex systems by nature, you will still have to deal with a lot of 3rd party software to make a viable mmo operate properly with unity
Yeah, I was hoping the RakNet update did more but then I saw this: http://unity3d.com/unity/engine/networking
ahh well maybe 3.5 or something.
for mmo you will always need to use something external in addition. the unity architecture is not even remotely usable for an mmo backend
I thought we established mmos were imposible to make!!
They are. You have to go outside and imagine a fully completed MMO.
But in order to do that you need to write a simple fantasy game and wait for it to become self aware.
Try as we might to tell people they have to start simple, nobody seems to understand why 
Nothing makes a MMO viable. A MMO is a massive project that needs a lot investment and manpower. As a rule of thumb: If you can’t run Disneyland, don’t expect to run an MMO.
I hope I can, Unity has sockets right?
OMG a disneyland MMO! Thanks thomasmahler!
All jokes aside… I don’t see many changes in Unity3 that make an MMO more viable… only thing that will make it viable is an interesting and solid design, motivated team… and a team with very intricate knowledge of networking, design principles etc (the list goes on…and on).
MMO a.k.a. More Morons Online 
That’s an interesting jab you are taking at the Unity team?
I’m curious how intricate would they have to get. Should they know the chipsets involved?
He means intricate on the end of the team making the final product.
Unity can’t make a superbly optimized networking system for 100s of players because it doesn’t know the format of what you are doing. The same reason they don’t have AI, in that it’s an area of a game that must be incredibly customized to the game in development. In those situations it’s up to you to either find the customized middleware solution that works in your case, or to create your own. If they made MMO networking, it’d be garbage for anyone making a multiplayer game with a low player count that requires fast as possible pings. Currently they provide something around the middle built in, with the possibility of extension, which is the easiest way to satisfy the most people.
Yes, of course. . .I see that now.
When you say “around the middle”, would a 100 simultaneous players be in that range?
I’d say not likely, but it’s incredibly dependent on the style of game, and your personal design. You could probably get 1000 people playing together… if it was a turn based game. But if you want a fast paced split second reaction FPS, you’ll probably want less than a dozen or so.
As you add players, latency will increase, and at some point it will become more than is acceptable for your game, the best I can suggest is to prototype something out and stress test.
Oh I’m not jabbing the Unity team at all, they are amazing. Unity is an incredible product (esp for free!). All I mean is that Unity’s goal is not as an MMO factory… instead it allows people to make whatever flavours of games they want. An MMO with unity is indeed possible and is an interesting proposition, but pulling it off will require very deep knowledge of how to build subsystems such as networking etc.
Unity is the tool, it’s up to the team using it to do anything of value. Just like a hammer doesn’t make a great carpenter or a guitar make a great musician.
Latency is the issue, isn’t it. And accessing sockets via C# probably doesn’t help much.
Unity is perfect for making a MMO client. All the other stuff that MMo needs will be the problem.
Knowing where to hook into and what restrictions are in place will be my problem.
Surely
Unity 3 has lots of features required for most MMOs, the most important of them is path finding. Because most MMOs have big scenes, oclusion culling might help alot as well.
Additional editor scripting apis and text support for doing merges in big projects are all features which are really useful in production of big games and MMOs are big projects most of the times.
However most of the heavy lifting in a MMO is on the server side and networking stuff and content creation.
We have an article on the more general question
Is it possible to make a MMO with unity?
And an article on Choosing your networking middleware for unity games
Holy necro-posting.
Ashkan, this post was 2 years old… and Unity 4 is now available. No sense in resurrecting obsolete information.