Compiling with RakNet Network engine
by Bobby Leighton · in Torque Game Engine Advanced · 09/03/2006 (3:41 pm) · 21 replies
Ok so I was looking around on the net and came across this..."RakNet" It list is capabilities with UDP and I was reading comments others had made,one had said this wouldnt be a bad platform lanch an MMO from...looking around further apparently this code is also used in the commercial game engine "Reality Engine" which If id had the money wouldn't have went with because of the MMO capabilities of the Engine out of the box...anyone think this would be worth mixing into Torque please speak up!!
Definatly need advice here....Im thinkingg my game will hinge on this...but as im currently starting College in video game simulation technology I might find better solutions in class(one of which is geared to MMO design)
and if any one is interested in getting good education in the making of these games let me know and ill give you my schools info(if youll take all classes online they start semesters every month almost and the take FASFA loans...great deal!!
Definatly need advice here....Im thinkingg my game will hinge on this...but as im currently starting College in video game simulation technology I might find better solutions in class(one of which is geared to MMO design)
and if any one is interested in getting good education in the making of these games let me know and ill give you my schools info(if youll take all classes online they start semesters every month almost and the take FASFA loans...great deal!!
#2
Disclaimer:
This is just my opinion. Other, more experienced coders may have other opinions, and that's ok. :)
09/03/2006 (4:52 pm)
Raknet is a great networking engine, but out-of-the-box, it isn't much better than the networking already implemented in TGE/TSE. There is one person (that I know of) that is working on a mod to extend it to work for mmg, but, on the otherhand, there is alreay a working MMG right here. I don't think the work to implement it into Torque would be worth it. Disclaimer:
This is just my opinion. Other, more experienced coders may have other opinions, and that's ok. :)
#3
[Ishbuu]
09/03/2006 (5:56 pm)
RakNet looks quite cool, i have o idea if you'll be able to combine it with torque, I think TGE's networking is fairly intergrated with its internals. (as in REALLY intergrated) but if you're feeling suicidal be my guest :)[Ishbuu]
#4
09/03/2006 (6:30 pm)
That's kinda what I though...i just haven't..(no offence Mr. Ritter)...Seen a lot done in the area of MMO except his and i really don't want to have to go Python,dont get me wrong I think it is a great language . I really just don't want to stretch Torque too thin...I would in other words, wish to keep with the C++ modifying the Torque engine in its native language and for smaller functions the Torque script seems more than attiquite....is it going to be that hard for me make Master-server program with C++ with the Idea of integrating it into Torque ?Can one be written using just script...im not too familiar with "script"..Played a little with 3dgamestudio and their scripting language but Torque won and here I am...I like to program with C ,and it is what I'm am going to school for so want to stay with it and use the next 3 years of school to help get me the precise education that i know this is going to take.
#5
09/03/2006 (6:34 pm)
You can buy TNL to use as a generic masterserver for torque. It's torques own masterserver from my understanding.
#6
09/04/2006 (4:44 pm)
Thank you...that would certianly would be a great place to start..do you think that I could perhaps get the structure of a server in one world to handle hundreds of thosands if not millions ,just in onw world.....I have looked at few expensive commercial packages and some of them boast this is in there networking package Masterserver...but i think I would look at the Torque one 1st....ty
#7
I'm sure TNL has an example, or at the least, the docs explaining how to make it.
09/04/2006 (7:25 pm)
I don't know. I haven't got that far with networking yet. The only thing I ever did networking wise, was to make a small chat program with raknet. It worked if you liked to draw your text with your mouse. :DI'm sure TNL has an example, or at the least, the docs explaining how to make it.
#8
You can download and experiment with TorqueNet in the GPL version to see what you need to do to make a MMO work. It will be a huge undertaking, larger than me as a one-man team making a combination of the artwork for Resident Evil 4 with the gameplay of Devil May Cry and the open-ended single-player gameplay of Grand Theft Auto. And it isn't because of the network model. You will have such huge challenges (and massive growth considerations) no matter what engine you use. Some may make some levels easier. Some may make it more difficult. And so you have to find what works best with your team.
Who is your team? What is your experience? How much can you pay to make your experience happen? Does your dream match our experience or the money you are willing to pay to make it happen? Are you motivated and experienced and dedicated enough to work with indie technology and stick with it unlike Duke Nukem Forever to make your title?
There are a lot of questions. It may seem like naysaying, but don't take it that way. Take it as questions that you need to ask and answer honestly. If your team consists of people who have used Photoshop a couple of times and Blender a little because it is free, but have no other experience, then you have huge elements to overcome before thinking about subscriptions and the mad cash involved.
I have no idea what your team is composed of, but no matter how negative I seem, take everything that I say as a positive view. I don't try to discourage. I try to make people think about things beyond the dreams. Dreams are wonderful. I love them. Dreams don't often get translated because dreams are imprecise and often bigger (or based on multi-million dollar budgets that have pronounced a mod recognition and the idea that one can make their own game which is on the same or higher level than the original--Counterstrike is an excellent example of this).
I want your game to succeed. But I want it also to be something that I want to play (my bias). And in my view no MMO so far has been worth the price of admission, let alone a subscription. Yet again, that's my bias. Look at why certain games work and others don't.
Sure, it comes down to the network, but it also comes down to gameplay. And most MMO gameplay is based on 1980's gameplay and banks on advanced technology.
09/04/2006 (7:42 pm)
Most that boast that also expect massive server and bandwidth requirements that are handled if available. If you have them available, and have the customer base (which is the key rather than the dream), then it is great. If you have an angel investor to invest in software, hardware, and networking infrastructure than GO WITH THEM. If you do not, then you will have to right-size your dream to the reality. What do you have in practice that makes you better than others who have it in practice?. People who have experience with proven engines have a bit of street cred, even if with the people who preview their works (IGN, GameSpy, GameSpot, etc). If you are not, you have a lot of ground to cover, and RuneScape gave people the free initiative. And others have picked up on that, and you have to compete with the many free projects that live and die monthly on the desire and whim of developers with dreams as compared to those which are semi-constant commercial ideals with huge budgets.You can download and experiment with TorqueNet in the GPL version to see what you need to do to make a MMO work. It will be a huge undertaking, larger than me as a one-man team making a combination of the artwork for Resident Evil 4 with the gameplay of Devil May Cry and the open-ended single-player gameplay of Grand Theft Auto. And it isn't because of the network model. You will have such huge challenges (and massive growth considerations) no matter what engine you use. Some may make some levels easier. Some may make it more difficult. And so you have to find what works best with your team.
Who is your team? What is your experience? How much can you pay to make your experience happen? Does your dream match our experience or the money you are willing to pay to make it happen? Are you motivated and experienced and dedicated enough to work with indie technology and stick with it unlike Duke Nukem Forever to make your title?
There are a lot of questions. It may seem like naysaying, but don't take it that way. Take it as questions that you need to ask and answer honestly. If your team consists of people who have used Photoshop a couple of times and Blender a little because it is free, but have no other experience, then you have huge elements to overcome before thinking about subscriptions and the mad cash involved.
I have no idea what your team is composed of, but no matter how negative I seem, take everything that I say as a positive view. I don't try to discourage. I try to make people think about things beyond the dreams. Dreams are wonderful. I love them. Dreams don't often get translated because dreams are imprecise and often bigger (or based on multi-million dollar budgets that have pronounced a mod recognition and the idea that one can make their own game which is on the same or higher level than the original--Counterstrike is an excellent example of this).
I want your game to succeed. But I want it also to be something that I want to play (my bias). And in my view no MMO so far has been worth the price of admission, let alone a subscription. Yet again, that's my bias. Look at why certain games work and others don't.
Sure, it comes down to the network, but it also comes down to gameplay. And most MMO gameplay is based on 1980's gameplay and banks on advanced technology.
#9
We in mind for the future(my partner has a degree in computer science...and I'm starting school that is endorsed by major game companies)Bought a HP Proliant with 12 gigs of ram 4 Xeon pross' and a crap load of hot swap drives...a half way decent router+a backup...a 3ghz P4 800meg FSB for zone server as well as several other computers to be used for login servers or maybe web server...Figured this would get us started and I've got a couple of production artist working in their spare time 4 me...and my self an accomplished artist(at least in my corner of the world)...I know i have a road ahead of me and that why the classes I'm taking focus on MMO technology and implementing it as well as other things I cant remember off the top off my head cause i just started,but i have a fairly good handle on torque but am waiting for MS 4 to really get started...most of what im working on involves lighting on the terrain since all we are doing is prototyping the code up and concentrating on the construction of the persistent world...then later add our models instead of a bunch of korks or soldiers...and put up our own buildings..
hope I've answered a few things for you....and also I figured id find some talent in projects with other class mates with the same goals...
also i appreciate the honesty most people wont answer like that they just fill you with BS...thats why I like you guys!!
09/04/2006 (10:45 pm)
Yeah my goal is to recreate some kind of Pen and Paper style D&D game with MMo tech thrown in...sorta....We in mind for the future(my partner has a degree in computer science...and I'm starting school that is endorsed by major game companies)Bought a HP Proliant with 12 gigs of ram 4 Xeon pross' and a crap load of hot swap drives...a half way decent router+a backup...a 3ghz P4 800meg FSB for zone server as well as several other computers to be used for login servers or maybe web server...Figured this would get us started and I've got a couple of production artist working in their spare time 4 me...and my self an accomplished artist(at least in my corner of the world)...I know i have a road ahead of me and that why the classes I'm taking focus on MMO technology and implementing it as well as other things I cant remember off the top off my head cause i just started,but i have a fairly good handle on torque but am waiting for MS 4 to really get started...most of what im working on involves lighting on the terrain since all we are doing is prototyping the code up and concentrating on the construction of the persistent world...then later add our models instead of a bunch of korks or soldiers...and put up our own buildings..
hope I've answered a few things for you....and also I figured id find some talent in projects with other class mates with the same goals...
also i appreciate the honesty most people wont answer like that they just fill you with BS...thats why I like you guys!!
#10
09/04/2006 (10:59 pm)
Did you say.. twelve gigs of ram?! Do you plan on simulating each blade of grass? 
#11
When it comes to world servers/zone servers, your better off using Torque's built in networking. It has been tried and tested, and performs well. It's really set up now though for fast action games, you could save some bandwidth and cpu power by lessening some of the strictness and update times.
Also I did want to comment on coding in C vs. Python or script. I started coding in 6502 assembly, and prior to the 68040/386 line of processors I was convinced that C and higher level languages were for lesser programmers. In this day and age though, and especially in MMOGs I would disagree. I have made my living the past 8 years in MMOG database related work and knowing a pretty good deal about how the majority of existing MMOGs work... I think the more that you can put into a scripting language the better. Script programming is pretty much the standard is MMOGs for a variety of reasons:
- Many changes do not require an engine to recompile or a patch to be sent. The more uptime you have, the more popular your game will be with players.
- It's easier to restrict employees access to things that you don't want them fiddling with.
- It's easier for new employees to pick up. Not everyone has programming skills, but an engine like Python for example is extremely easy to learn.
09/04/2006 (11:36 pm)
I used Raknet a couple of years ago, and while it is an engine that is capable of creating an MMOG, and does have some features that are pretty useful for that, I don't think it would be worth the effort of porting Torque to use it in general networking. What it would be useful for though is to use it for the backend stuff, to route players to world servers, authentication and patching, things of that nature.When it comes to world servers/zone servers, your better off using Torque's built in networking. It has been tried and tested, and performs well. It's really set up now though for fast action games, you could save some bandwidth and cpu power by lessening some of the strictness and update times.
Also I did want to comment on coding in C vs. Python or script. I started coding in 6502 assembly, and prior to the 68040/386 line of processors I was convinced that C and higher level languages were for lesser programmers. In this day and age though, and especially in MMOGs I would disagree. I have made my living the past 8 years in MMOG database related work and knowing a pretty good deal about how the majority of existing MMOGs work... I think the more that you can put into a scripting language the better. Script programming is pretty much the standard is MMOGs for a variety of reasons:
- Many changes do not require an engine to recompile or a patch to be sent. The more uptime you have, the more popular your game will be with players.
- It's easier to restrict employees access to things that you don't want them fiddling with.
- It's easier for new employees to pick up. Not everyone has programming skills, but an engine like Python for example is extremely easy to learn.
#12
I will take you advice on scripting though...It was and is now confirmed by you that Scripting is a solid method for writing in-game functions like Spell casting or the way damage is handled...i like Torque and from what I've been seeing there arnt many Engines better than Torque!!Ive looked and even tried a demo from Epic of the latest Unreal Engine (although AMAZING) Torque is on its way to being just as good an engine ...just as feature rich...and the network code is probably a little better I notice that with out modification they only support a couple of hundred player at a time on Unreal..I'm sure this can be changed but if you have to Mod an engine to get what you want...then why not spend a couple hundred instead of a couple of thousand
AND yes 12 Gigs or ram....see servers like the one we bought handle up 16 gigs...some can handle much more!! and they are 4 @900mghz each...windows server 2003 ...dont know the FSB at the moment
09/05/2006 (9:20 am)
I have played with Python....but they arn't teaching it in my school and I don't know how they would react to me using code that differ from what they are requiring of me...However I like Python and have followed Josh Ritter's work as closely as I can, and I must say the stuff he has merged with to make "MOM" is AWSOME!!However, I will never get my partner or many other people on board for this.I think their thinking is that it is not a standard language that their familiar with and to add another one into the mix when its not necessary and when its not native to the engine kit could potentially cause more problems that will have to be fleshed out later and/or worked around in some fashion...something I'm not a fan ofI will take you advice on scripting though...It was and is now confirmed by you that Scripting is a solid method for writing in-game functions like Spell casting or the way damage is handled...i like Torque and from what I've been seeing there arnt many Engines better than Torque!!Ive looked and even tried a demo from Epic of the latest Unreal Engine (although AMAZING) Torque is on its way to being just as good an engine ...just as feature rich...and the network code is probably a little better I notice that with out modification they only support a couple of hundred player at a time on Unreal..I'm sure this can be changed but if you have to Mod an engine to get what you want...then why not spend a couple hundred instead of a couple of thousand
AND yes 12 Gigs or ram....see servers like the one we bought handle up 16 gigs...some can handle much more!! and they are 4 @900mghz each...windows server 2003 ...dont know the FSB at the moment
#13
I think you put the cart before the horse was even invented.
I'm working on some research for an MMO framework this year (I've got a research year) and one thing I can see already is that if you really want to scale something, then Torque and its very good networking, really isnt the thing to use.
Torque's object model is far too tightly coupled with its render model and in general isnt as clean and extensible as it needs to be for a MMO (imho). So essentially I plan on rearchitecting some of the internals of torge to get a cleaner run at the networking issues.
I'd suggest you do some reading on MMO architectures, design, etc etc. Buying a server seriously was the last thing you needed to do.
Myself, I've got no ideas outside of my own scope. My "MMO" will be running on cheap ass hardware and hosting maybe a few hundred users, not the thousands that commercial MMO's are doing per server farm. I just like experimenting with scalability and online persistant worlds..
09/05/2006 (2:47 pm)
Why in gods name did you already purchase a server when you have really no clue what it is you are going to make or how?I think you put the cart before the horse was even invented.
I'm working on some research for an MMO framework this year (I've got a research year) and one thing I can see already is that if you really want to scale something, then Torque and its very good networking, really isnt the thing to use.
Torque's object model is far too tightly coupled with its render model and in general isnt as clean and extensible as it needs to be for a MMO (imho). So essentially I plan on rearchitecting some of the internals of torge to get a cleaner run at the networking issues.
I'd suggest you do some reading on MMO architectures, design, etc etc. Buying a server seriously was the last thing you needed to do.
Myself, I've got no ideas outside of my own scope. My "MMO" will be running on cheap ass hardware and hosting maybe a few hundred users, not the thousands that commercial MMO's are doing per server farm. I just like experimenting with scalability and online persistant worlds..
#14
09/05/2006 (3:01 pm)
Aye, I too found the server comment funny. You wont need that much RAM for a Torque server. A database is another story, but still - I would make my game first, then adapt the server to what I need, not the other way around.
#16
Phil gave some helpful advice. You better get used to that sometimes people will give you information that might not always be what you wanted to hear. If you cant stand that then I honestly dont know why you were asking in the first place.
09/05/2006 (4:17 pm)
Question is if yours is any more helpful?Phil gave some helpful advice. You better get used to that sometimes people will give you information that might not always be what you wanted to hear. If you cant stand that then I honestly dont know why you were asking in the first place.
#17
We just ordered a couple more servers and will soon be running on 5 dual Xeon/Opertons... We've also started phase 3 of the server topology upgrading... you can check out the results of phase 2 here: http://www.garagegames.com/blogs/4280/10708
09/05/2006 (4:48 pm)
We developed MoM through alpha and beta on a 2100mhz Athlon server with a 512k DSL uplink :)We just ordered a couple more servers and will soon be running on 5 dual Xeon/Opertons... We've also started phase 3 of the server topology upgrading... you can check out the results of phase 2 here: http://www.garagegames.com/blogs/4280/10708
#18
09/05/2006 (5:30 pm)
The only reason we bought anything guys is because we had the money at that time....if we had waited I would have spent it on Video games and God knows what else!!LOL....but as far as why we bought sooo much?Well wouldn't you if you had the chance?We wenrt going to have the money forever so we got this in hopes to be able to work on a large scale server setup and have net servers to get online business of other types for their respected web pages...so its not ALL for my Torque game...but if I was gonna buy another engine that was more suited for MMO's I would have to spend a lot more to even come close to the over all graphic quality of Torque...I believe an maybe I'm wrong...but I believe Torque is way scalable and with the correct modifications I don't see how it(especially with the price),how Torque isn't suited for all genres regardless of it out of the box setup?Also for MMO example of a shooter Type engine...which out of the box is what Torque is...Take Unreal 2...Lineage 2 is the game there handle more per server than the engine was designed to take...I just want to do similar things with Torque...I know it seems like a lot of work but I have 3 years at a school that I'm taking all online by the way...and wanted to have some of the equipment at my disposal that their teaching me on...hope this helps....
#19
In general, if you are going to develop multiplayer games, having a smaller test machine is fine. Dont buy pretty expensive server hardware up-front when you really dont have any clue of your bandwidth needs or architecture.
As Josh's example points out, you can develop with a low-end machine and scale up. By the time Robert gets his game to a useful state, his server machine will likely be obsolete, he would have been better off to wait and buy a top-spec server machine once he had a clear idea of his requirements.
But its hit choice. I wanted to point out that I would do differently is all.
09/17/2006 (7:03 am)
Mike, I was just kind of pointing out that getting a server was a premature thing to do. Think of it as general advice for anyone else thinking of doing the same rather than directed at Robert.In general, if you are going to develop multiplayer games, having a smaller test machine is fine. Dont buy pretty expensive server hardware up-front when you really dont have any clue of your bandwidth needs or architecture.
As Josh's example points out, you can develop with a low-end machine and scale up. By the time Robert gets his game to a useful state, his server machine will likely be obsolete, he would have been better off to wait and buy a top-spec server machine once he had a clear idea of his requirements.
But its hit choice. I wanted to point out that I would do differently is all.
#20
However I should point out that It was not my money that bought this it was an investor(partner) and we do more at the moment that make online games...we are creating a data base for ...well another couple of projects....
you should not go out and but top of the line equipment when you haven't made you game yet...we bought something old...I should point out also that even thought the server has 4 Xeons..they run on a FSB that is only 100MGhz....pc100 ram.....we dropped some cash...or Rather my partner did...but it was only a couple of thousand...not much compared to whats ahead...all in all with the 2 servers,.. big one and a small dell to handle zoning we spent about 2,000$...that not that much for testing equipment is it?
Im not trying to be difficult but i like this line of conversation:)
10/02/2006 (4:06 pm)
Phil is absolutely correct with out a doubt!!!However I should point out that It was not my money that bought this it was an investor(partner) and we do more at the moment that make online games...we are creating a data base for ...well another couple of projects....
you should not go out and but top of the line equipment when you haven't made you game yet...we bought something old...I should point out also that even thought the server has 4 Xeons..they run on a FSB that is only 100MGhz....pc100 ram.....we dropped some cash...or Rather my partner did...but it was only a couple of thousand...not much compared to whats ahead...all in all with the 2 servers,.. big one and a small dell to handle zoning we spent about 2,000$...that not that much for testing equipment is it?
Im not trying to be difficult but i like this line of conversation:)
Torque 3D Owner Marcus Daw