How to create a MMORPG?

I am not doing a add i just helping i think you will need go to Runscape in google and after i resgister you can see tips for your game.
Bye
Enjoy :wink:

 

making a MMORPG is a undertaking task, reason is it takes some time to create assets, which in the case or mmorpg means you have lots of open spaces to fill, that is why it takes years to fully develop one.

now, this is assuming you already have the engine and network done and game play logic scripted

if not then this may take you another couple of years to achieve.

I remember back in the days one could not even think of posting a thread on any game dev forum about how can i create a mmorpg. they will bash you, slap you on the face

well, those days are over.
making a mmorpg now adays is very doable, even for a single or a small team of indies.

there is lots of things you can get help from to speedup your work.

- why reinvent the wheel if there is one already working. will save you years of developement.

-assets, if you are a single dev, you can always buy already made assets packs, will save you even more time, months etc..

I think one does not have to wait for a fully created game to publish, this will save you another couple of years.
most of the mmorpg nowadays are always if not constantly updating the game by patching it, adding new features assets etc..

so , lets say, if you work smart from day one you can have a playable mmorpg before yo know it and keep on adding more stuff to it as it grows.

lets start from the beguining:

engine to choose:
- Bigworld Tech. for $300/year + royalites if a commercial game for up to 25 seats.
BWT is a fully mmo engine, designed specialy for it, you already have a AAA client and server done for you.
by far for my point of view this is one of the best if not the best networking server architecture for mmo outhere. to start, the server that will help you grow and scale up the user base with no special needs or tweaks, the server can handle millions of simultanious player connections.


-HeroCloud. HeroEngine indie is another good option, I personally havent used it, but looks another great option for creating you mmo game.
HE indie is priced at $5k license for up to 25 seats
+ yearly subscription and royalties, as per its site it looks it has quite some good tools that will help you speed up your work.
even HJ mmorpg game reference scripts to help even more speed up proccess.


the above 2 are out of the box engines specially designed for mmo games.

now there are some more alternatives, like Unity + photon, or smartfox server, or badumna.

-T3D with mmokit addon a proven indie mmo architecture. you can always tweak existing code and server and get what you need, not sure if it can handle massive players but looks that it can handle hundreads if not thoulsands if handled propertly distributed.

good thing about mmokit for T3D is that you alreay have all game logic to start with. just add you assets and create your world.


-ryzom open source mmorpg engine.
thsi is also a great option, you pretty much got everything there awell, game logic, client , server already made for you.
ryzom is another proven mmo engine that can handle masssive ammounts of players, the only drawback is the lack of tools and documentation to get you started. and if you dont mind the actual license agreement then you got yourself a great tool there.


independantly on what you will use, if you use an already developed engine tool as described above, then you are half way there, since this will save you years of developement.

also a must option is to create an online auto patcher, this way you can easily add more game stuff to your alreay running small scale initial game.

you can always start by adding your base rpg game logic a small well detailed equipped zone and a few quests and you alreay have your mmo game online running in no time.
then as you go with the autopatcher feature you will keep on adding zones, quests game logic etc...

the good thing is that you alreay know that you already have the backed networking architecture that will support more players as you player base grows which is the most important thing here.

I am pretty sure i missed lots of other options but this will at least get you some startup references.

 

We already know that develop a commercial MMORPG may take years of dev work, but the thread is about how to create one, or at least the base of one. If it will take years or not is not in vain, the question is how to develop.

I made some research and I realized that actually there are some interesting open source projects like planeshift, tibia ot, the mana world, stendhal and others. Some of these projects are really small, but has the main elements that a basic MMORPG should have. I am taking a look on these projects, maybe I post something in this thread in a near future.

 

As far as database vs flat file. Honestly, you should use both. load a flat file into a database when a character logs in, when he logs out (or loses connection) move the flat file to a backup location, then save a new flat file back into it's location. Very low processing time in the end, you get the stability of a database, longevity of a flat file. You can even create flat file logs. Say keep a flat file for a week, then dispense of it. If they haven't bitched in a week then they shouldn't really need it.

As far as computing cycles for databases. I am a db administrator, and have found MySql to be alot slower then MsSQL. Oracle is very fast.

I can't attest to various open source NoSQL servers though, but it could possibly be a different approach. (or, quite even what Blizzard uses) I would imagine that you could simply use 10-100 servers that all run the same world with millions of users on at the same time with the technology. However, this would range far beyond the point of this thread)

I have an interesting theory, though not complete. Something called a Cloud MMO. Where most of the processing is done by the user's computer rather then a server. Say John is fighting a monster. Well, John's computer has control of the monster, it attacks, he attacks and the like. Jim comes over, and asks (whats in the vicinity) The server says, John, and this monster. Both are being controlled by John's machine. Jim attacks the monster so that John wont die. Jim then interacts with John's machine rather then the server. All the things are handled by John's machine and the only thing that the server receives is location, hp, target and other minor things.

Say John gets tired of his fight, and simply turns his client (or machine) off. (bad John) Well, the computer never registered John receiving XP for killing the monster. It did recognize that John was the first person to attack it. It knows where John was and what his hit points were and who he was last attacking. However, the server would then say, wait a second, John is no longer attacking this person. Do I have another target? (poor Jim) And transfers the monster to the control of Jim's machine. If, however, Jim was not around, it would simply return the monster to it's original location and reset it's HP.

Database/Core wise, if you ran a osx, windows or linux/unix server with the server content on top of a database. You wouldn't need to run a copy of Unity3d at all. The engine would simply say, this monster is in this area, you do all the work, just give me back the results.

The theory says that the server doesn't have to calculate hits or damage but simply lets the client do all that. After all, the server has to maintain a database. Of course, its all theory.

 

The path to take for an indie MMO is to aim for a small number of players, perhaps a few thousand, and do something different.

This, I feel, is where indie developers have major advantage over the big publishers. These guys are too scared to lose money on a new idea, they have to spend 10's of millions almost by default, and again almost by default players expect more from a big AAA game. An indie MMO could potentially be very successful with a tiny budget, a crazy idea, and a small playerbase.

A couple of other things.

Have a look into Amazons EC2 cloud stuff. I've been looking into this and there is a huge potential there for indie developers. (Although you might want to shop around for cheaper options.) You can run a small linux server for around $60 a month, or a small Windows server for about $80. The caveat here is that the small sizes might not be useful for anymore than 50 concurrent players (I haven't tested it yet.)

However, it's terribly easy to scale up and create bigger instances. The beauty of this, is that you can scale as you grow. Start small at cheaper cost, and pay more as you get more players. Not many dedicated hosts offer that. It's worth checking out.

As for databases. I've downloaded Sql Server 2008 R and am evaluating that. It seems promising as it offers asynchronous processing which will help with scaling. The downside, being free and Microsofts equivalent of the Express versions of its IDEs, is that there is a 10GB limitation on database size.

Again for a small player base that might be all you need. Or, there might be ways around that by using multiple DBs for different tables.

But again as you grow your playerbase, you should be able to afford the additional costs. And interestingly, Amazon EC2 offers SQL Server instances that a fully licenced... no limitations.

Hope that helps.

 

I know of one MMO tutorial, put out by IBM in 2008. It uses Torque as the game engine, but many of the principals can be re-applied.

Have a look:

http://www.ibm.com/developerworks/library/ar-powerup1/index.html

Cheers,

Galen

 

Ok, now i've read some more, a couple of thoughts on cloud... Yes, cloud offers the kind of dynamic ramping of virtual infrastructure that something like an MMO would need to better align costs to actual usage (ie, you don't need to build out a server farm in advance of load). And there is at least one vendor specifically targetting this market (RightScale). Some technical things to consider on the sever side(as part of the overall software architecture):

As I recall, EC2 images can come in 32 and 64 bit instances, for DB a 64 bit instance (with, oh, 16 Gb of ram) can handle a fairly large load before you need to spin up more instances. But, you need a 64bit DB or it won't leverage the hardware anyway.

Someone mentioned Java, which (despite some claims) is still really a 32 bit environment (ok, the caveat is the 64 bit JVMs are still only using a staged 32 bit GC, which runs slower in 64 bit than you'd get if you just ran in 32 bit environments... trust me, I tested the h*ll out of all the JVMs available last year). IF you use Java, you really are better off using as small a heap as you can manage, and spawning more instances to scale (512 is a good number, but 1Gb can perform, just don't break 2Gb... IBMs JVMs (and all "stop the world" based GC implementations will slow considerably, and the concurrent GCs will start becoming unstable after sustained load).

In other languages, couple of Linux notes: like all Unix types, there is a default cap on RAM per process (AIX defaults to 2Gb, I think linux is the same). So if you're going over that remember to change the setting. However, Linux (interestingly) manages to fok processes about as fast as it spawns threads, so you're better off keeping each instance in a more manageable heap, and spawning additional processes there as well.

Windows... you're on your own MS tells me "it's all good". My experience says, the same process limitations still apply by default, but you need to account for slightly higher process spawn time to do this kind of verticle load distribution. (doesn't matter to a single user, but at MMO volume loads, 3 seconds can be the difference between my players are happy, and my MMO is down )

Also, with cloud "marketing" being what it is, you should know, that, while yes multiple VMs on a single server have an virtual network "across the bus" latency advantage, VMs on different servers (or worse, VMs that span physical servers) still suffer the same network latency as any other hosting solution (with the later encountering issues in-OS, much less between runtimes). So smaller VMs on bigger hardware is the way to go (not that you get to pick... that's the cloud "advantage"

Just some thougths,

Cheers,

Galen

 

I was curious as to a thought on a server/client using Unity3d for both. All data is stored in XMLs (character status, AI behavior/drop algorithm, etc.).

How many people would you propose could populate the world? (I'm guessing about 30ish)
How heavy would it be to host the server on a rather powerful machine with a lower end bandwidth limit?

Just curious. I know to have a more realistic amount of players inside one world, it would need significant scaling. But just wondering based on assumptions.

 

Read the other thread I will just add it and save my fingers.

http://forum.unity3d.com/threads/69592-wantz-mak-MMORPG/page1

I was discussing Scalability,

Unity as a server may be able to handle 50 to 60 if (a biggie) your server code is really simple and has very little AI, The server would need to disregard the graphics end and just run the JavaScript, c#, boo or whatever you program it in. Since you're going this route and having to program in all the logic, database access and all that why not just use a prebuilt server that you can expand on?

Smartfox pro cost money and is limited if you use a free version (20 CCU). You can't really see the potential of this new technology. Of course a lot of the "web" stuff is there for buddy list, rooms and so on but the real potential is just locked away until you spend some dollars. Not only that but all the scripting is interpreted with actionscript.

The SGS server http://sourceforge.net/projects/reddwarf/ (Sun Game Server, Darkstar) (aka RedDwarf) is ideal for a learning environment and interfacing it to a Unity client is not that difficult. It is not like RyZom where you need to have a masters in computer science and use C++ to compile and debug it until your eye bleed or spend 10 hours configuring it for a 13 servers to cluster. You're not going to have to learn server infracture and all that but you need a basic understanding of programming in Java, a little elbow grease in learning the API (which you also have to learn in Unity anyway and is the same as expanding classes in Unity) so you may as well get a free unlimited use copy and use it. This server is designed to make its own database for game play and keeps it persistent. Most of the real work in making a mmo server is in the background and already done for you. All you need to do is throw in the logic and gameplay of course you won't have account management but that should be another server running PHP5 or even java. You can still use XML files to load the players data or go with a real database backend like Jdbc driver. All this stuff is free and you can always get help in setting it up off a forum. If you rally want to make a MMORPG stop wasting your time with all the crap and use the right stuff.

Scalability of this server without multi-node.

http://sourceforge.net/apps/phpbb/reddwarf/viewtopic.php?f=1&t=501

Multi-node is the big factor in making this server a distributed programming high scaling server capable of some incredible stuff as I talked about in the other thread.

Making a MMORPG is not an easy task but the RIGHT middleware for the game engine you use will make it a LOT easier to develop one. The biggest cost is paying for time.

 

Not making an MMO, this is just a project my team and I have decided to undertake for school related reasons. We won't be expanding it to any backend server unless there is a reasonable interest from the players or we receive financial backing. I guess we're trying more of a network benchmark using equipment that you're average beginning indie dev will have (using a dedicated computer, not a server).

But thank you very much for an explanation. I imagine ours might be a bit heavy, we're geared towards implementing real time attacking (like spam "x" to spam the basic attack).

 

SGS sounds VERY fascinating, could you give us some more info? Such as, how much does SGS do for you in terms of coding? Does it have headless game engine features yet? (physics, adding scripts to interactive objects, performing all "Update" functions once per frame, for example)

Also, where can I get some n00bish info on SGS? I've found a very technical description in the red Dwarf docs, but nothing for the total noob. In this way, SmartFox outdoes red Dwarf.

 

SGS is actually called RDS now or RedDwarf server.

It is highly technical though and not very noob friendly. There are no layman papers so you're stuck with the engines class reference, The internal structure takes care of the networking and the game persistent database. You just pass it the tasks and the schedules and it does the rest. The design is to reduce the development time of making a network infracture (clustering and internal communications of multiple servers) not making the game so you still need to develop all the modules that a MMO would require.

I found out yesterday (through experience) the RDS is not production ready at this time so some features are broken. One of the big ones is server restart does not work as intended and the database is corrupt.

There are issues to work through right now so if you want a MMO up and running in the next several years you probably better off going to a known working model.

 

So, to clarify, use SmartFox? Or something else?

 

It is up to you on what engine you use. No matter which one it is you're going to have to put forth effort.

IMHO - start by working through the game design, make an outline of the requirements and then choose the best engine that can do what you intend to is the right way to do it.

 

If this is true the game will create itself.

 

I would say there is another option as well you could use ElectroServer with EUP from Electrotank. It contains many of the concepts out of the box for an MMO style game and is easily extendable.

 

Demi, could you compare ElectroServer and SmartFox if you know anything about them?

littlelingo, I don't beleive EUP works with other engines.

 

+1
Even when designing single player games it takes a lot of time and efforts to design at least playable, if not completely new gameplay. In MMOs though, you have to consider thousands of people potentially exploiting your gameplay mechanics in their own kind. And thats just technical aspect. Another thing is uniqueness of your gameworld setting. Consider that,post-apocalyptic-zombie-eating-orc-spaceship-pilots have already been done. You need something new to shine through hordes of MMOs. And at last, ask your friends their favorite game. One will choose chess, another - checkers, tic-tac-toe, rock-paper-scissors, Cod:MW, etc..And combine them together, let them play. If they will say it was ok, they are lying. If they say it IS great, then you can either develop MMOs or get a degree in sociology.
On a serious note though, it's ridiculous to talk HOW, while you're not sure WHY, and will customers even take look at your game. And it's possible to make "another" MMO successful with technical aspect only, but it has to be something unique and enjoyable by the thousands. And is it worth the time?

 

Sorry, I have no knowledge of ElectroServer so can not make a comparison.

 

Not really, at no moment I said that the brain created itself.

 

It depends what you mean by other game engines? EUP is an extension that runs on top of ElectroServer (and currently has a C# API) so it is compatible with Unity or many other C# implementations (XNA, Ogre, etc.). If there would be other APIs that are needed to be supported in the short term, that could certainly be discussed but IMO is out of scope for this thread. ElectroServer has several APIs (C#, Objective-C, Java, JavaScript, AS3) that would tie nicely with a large number of game engines out there such as Unity, Ogre, etc.

As far as direct comparisons, there are several threads out there that goes into detail about both. Here are some ES links that go over the features, etc.:

http://www.electrotank.com/es5.html
http://www.electrotank.com/electrotank-news/54-es-news/296-case-study-why-aquiris-and-aeria-games-chose-electroserver5-for-their-unity3-fps.html

HTH,

-- Clint

 

so.... i have a fully completed rpg ( massive world completed, quests, monsters, monster drops, items done gui done and everyting) , now what are the elements i would need to turn it into multiplayer, say if i only want 200 people per server,

i dont have any experience with multiplayer stuff, but i can code other stuff to make the game.

 

Unfortunately, coding for a morpg is very different than a normal rpg. You have to design and build your game from the ground up with ElectroServer/SmartFox in mind.

 

grri hate this networking stuff -.-i dont get it at all.... a small tut would b nice to explain what everyting means

 

I definitely agree. PM me with ya questions, I'll do my best to answer them.

 

ok for the first question, i have created all my scripts, so what in general would i need to change in them so that they are multiplayer compatable, for example i have an mob spawining script, attribute script ( character stats) etc etc, so on scripts like that what do i have to do differently to make them compatible with multiplayer

i also hear alot that unity cant handle big levels on multiplayer, but my levels are actually prety small but they are stiched togeather with portals, meaning that at the end of every level there is a portal that then commands unity to load up the next scene ( which is actualy the next part of the level) so having my world like this does it help?

 

Dude, I said Private Message me. (PM)

Oh, and to answer your question, you would need to use ElectroServer anyway, not Unity Networking, so not a single line of code would remain the same. Sorry, but you have to start from scratch.

 

@ Quietus u are highly immature

@ MTracerStudios ok, thnx for reply next time i will pm u, btw this sux -.-

 

Quietus: You are like justice. Blind but fair. Some times I read your posts and go "respect". Other times I go "ouch". So whenever I see you reply to a post (especially one made by me) i always hold my breath before reading it... Is he going to swing in my favor or not?

Anyways, your post may not be very helpful to noobs, but having struggled with the same question recently, I can see where you are coming from.

I had a number of scripts that do some rather intricate tracking and linking etc and they work quite nicely... then when I tried to go multiplayer I noticed that dropping the network stuff on the object causes certain stuff to be duplicated over the network that simply shouldn't be. For instance, imagine every player's MainCamera being transmitted over the network... The Sky has never looked so weird.

But then there is the singletons that I use as static, non-destroy objects. Making a network game where all players use the same global settings like "PlayerName"... Just not on. So yeah, making an entire game first and then thinking "how can I make this multiplayer, now?" required that I change the entire working of the game...

With each script and component only announcing it's error after the fact, unless you know exactly what you are doing when and where, you would in fact be in over your head trying to convert it.

Keep networking in mind from the start? I have to agree with that!

 

It seems to me that many people just want to know where to begin. That seems to be the original purpose of this post. I've seen many posts where someone will list off tons of things needed to get an MMORPG up and running but then never are able to elaborate. Even those that claim to have experience working on this style of game. As one poster mentioned a working PLAN would be great. BUT...

I think if a 1-2 person team were able to get enough of a game created, even using Unity and free 3d models, etc., they may be able to get others with more specialized skills to take interest in their project and contribute. Even if someone can't "join" a team but helps design a "monster" or something, it's a start. Other than that if the game is fully realized on paper and a semi-working model then it's just a matter of getting a programmer/3d Designer/ Networking engineer/ etc.? Whoever... on board. I'm not saying that is going to be an easy task but if you have something to show someone it is far more likely that they will take you seriously and want to join. Hell, even if you just have something limping along, anything tangible is more likely to garner interest than a bunch of scribblings in a notebook.

I have a buddy that helped design/program a full mmo for kids http://zoodlezone.com/ They had a lead developer, 2 programmers, 2 testers, and 1 network guy. Other than that 1 secretary. It's not the best game in the world but it's pretty good... point is THEY DID IT.

If you live near a University/College and have a in-depth plan and a working or semi-working model it may be possible to recruit students to do free work for their resumes and portfolios. I would have to imagine there are a few there with similar aspirations. Just saying you built an MMO with a small indie/hobby team has to help your chances getting a job vs. someone with no experience, even if it's a crappy game... it's a crappy game that works! Otherwise, use your crappy game to recruit quality people to "remake" your game into something more beautiful.

 

What I'm doing (and i find its working so far) is to plan EVERYTHING you want in the game, and everything you want it to do at the surface

then plan what it needs to do under the surface
(take as long as you need planning/designing)

once you have designed everything you want in the game start making it but only include what you have designed/planned, so you don't get overwhelmed.

I'm aiming for a MMO running on SFS made with Unity using MySQL with around 100 players max.

 

I should note that for a long time in the Asian market, there are MMO's that are completely match-based (persistent character data, subscription or microtransaction based revenue, no massive world but matches in instanced maps), and they survive in the market just as well.

 

I've seen horrible sounding games raise over $120,000 on kickstarter and the like.

If someone were to make a very fun and innovative MMORPG, I see no reason why they could not raise an enormous amount of money to finish it or add a great amount to it.

After all... if weird, crappy sounding games can raise scores of thousands of dollars in a few months, I see no reason why a fun, successful, popular, innovative game wouldn't. Especially if people can play it and think "Wow, this is fun!"

 

I like Photon,maybe take a try..

 

Just curious :
Are MMOs only games that are monster/alien/zombie versus human? I know that they're also Massively Multiplayer, but are they uniquely that type? Or would a vehicular multiplayer game qualify too? (I'm thinking "Mars Explorer, Mars Explorer")

 

Probably, take a look at test drive unlimited. That's classed as an MMORG (Massively Multiplayer Online Racing Game).

 

MMOs are games that are massive multiplayer online.

 

Like Eve Online? Because that is definitely an MMO.

 

I think any game that goes beyond 64 players (or whatever the new maximum for most games is, as I've heard of a few recently that mightve broken 128 players in one battle) is realistically a contender to be categorized as a MMO.

Why? Because there is little actual categorization as to what "massively" means. Since even the largest FPS game does not go beyond 64 (or more recently possbily 128) players, I believe that anything above those numbers is quite massive. 128 alone is massive. 64? It's pretty large, and IMO just as massive as a MMO.

My justification is that in most MMORPG's, even the most popular, there are actually rarely ever more than 5-30 players in any one area at any one time. Even in massive areas (WoW Capital Cities) having more than 10 players TALKING in global chat is a big deal.
MMO's are NOT millions of users online at once. MMO's are at most thousands online at once, spread across large worlds, separated in level brackets, and centralized in only a handful locations (if any).

Even in the most populated server on WoW, you do not have millions of players. So I think anything above 64 players in any one area, and in any type of genre, is quite MMO, at least in my opinion. Whether that is racing, FPS, or RPG.

 

MMO does not simply entail the number of players. The "Massive" part of MMO deals with the size of the game world. A game such as MAG where you have 150 people playing on a relatively small map is NOT an MMO. the game world itself must be massive in size and a single instance... no map based playing (not taking into account instances inside of the main instance such as WoW does). a normal MMO server supports thousands of people at one time on the same instance.

 

Ok, thanks everyone! This has helped me understand what an MMO really is.

Thanks again!

Arthur

 

I would argue that MMO is still loosely defined, and game designers have a great deal of lattitude and still be MMO in the end. Wikipedia has the following definition:

in this case I would say, MMO by definition is Multi-player, has a "online" or internet hosted element(s), and scalalable (capability). Nothing in that description (or any other I've seen used in the industry) says all players must be "present" within the same world, at the same time. Nor must there be a "one world", just the notion that the world or worlds are persistent (that doesn't mean they necessarily evolve overtime, they just don't pop in and out of existance on you.

There's a lot of room for games that can validly call themselves MMO.

Now, WoW is the 800 lbs gorrila, true. And it certainly has a whole host of criteria, if you were to set your sights on "only" WoW as a model. Of course, despite all the marketing, I've never seen anywhere near 12 Million people while playing WoW (in fact, anything approaching 4-6 dozen "visible" other players begins to cause stuttering in my game play).

I'll add one other thought, people are really hung up on the technical skills / approaches necessary to actually create an MMO. and yes, it's a significant non-trivial task to do any of the technical pieces, so don't misunderstand me, if you can get to "it's done" you have already crossed the massive mountain range. But once done, a successful MMO isn't just going to deploy to a "cloud", spit the game out in Steam and let the money roll in. The business and ongoing operations of an MMO is the longer term, and possibly harder beast to tackle... a beast very very few of those who do make their game have managed to make work... yet.

As I said in another thread, can you make an MMO? Yes. (I'll add) can you successfully run a MMO business? Yes.

It's more probable though, that you'll tell your first AA meeting "my MMO sent me".

Cheers,

Galen

 

You seem to have misunderstood my description as it is the same as the one you gave.

There CAN be multiple servers. Each server would have one instance of the world where all players thrive. Not all players would be on the same server, of course. Not all MMOs need to be an RPG like WoW. Almost any genre of game can be made into an MMO.

 

The true definition of an MMO

"Something we play, and they make"

 

List of Free 2 Play MMOs:

http://en.wikipedia.org/wiki/List_of_free_massively_multiplayer_online_games

List of MMORPGs:

http://en.wikipedia.org/wiki/List_of_massively_multiplayer_online_role-playing_games

 

Yup. I'm making an MMORPG, then I'm going to publish it, make a huge loss in the first year of it's release, then sell out to Disney

 

You've just unlocked the 2 year necro achievement!

 

"achievement unlocked!"

 

Great! When do I get the Bankrupt Game Designer Looking For Work award?