How to fix the Multiplayer in Demigod

By Ironrock on November 22, 2009 11:17:04 PM from

HOST DEDICATED SERVERS

Have lower to higher RANKED SERVERS. Have servers evaluate player rankings and have dedicated servers reject over-ranked players trying to join lower ranked servers and vice versa. Players would have to "fight" thier way up to the higher level servers. This would "BALANCE" games. Players would always be playing people within thier skill sets.

The dedicated servers will keep cycling maps with varying amounts of slots available..always maxing the slots. (This meaning if it is a 5v5 map...10 slots open and and will start once filled. No player is in control of slots. Nor can they see what demigods other players choose. They would only see names and rankings.

If players get over a certain threshold of disconnects, they would only be able to join the lowest ranked servers until thier threshold is back within tolerance.

There are a multitude of ways to make this happen with cost in mind.

1. Have dedicated servers in different geographic areas.

2. Have each server with slots and a que.

3. As a game progresses and some one "rage quits" or disconnects..the slot is open and a player in the que with the same demigod is brought in at the same level and items.

4. Disconnected player gets a loss.

5. New connected player gets neither win or loss but gets exp and favor.

6. Players that leave cause no desync issues with others as there is no peer to peer involved.

7. Custom could be hosted peer to peer.

Add dedicated "Clan servers" in which players that form a Clan can rent. Minimum Clan size and upfront fee for X number of months (I would say 6 mos to a year). Dedicated admin for each "Server" (Server would be based on total slots hosted with LIVE players concurrently). Clan ranks would be captured from these servers based on minimum settings defined by Clan community or SD.

Final note: I know the peer to peer methodology is for cost savings. I would say that there would be community interest in hosting dedicated ranked servers. This could be implemented as a beta program and I believe that you will find it more successful than you might think.

0 Karma | 58 Replies

If all you're concerned about is which machine is calling which when the connections are set up, the proxies already abstract that part. Still doesn't make the game not peer-to-peer, though, just easier to establish the connections.

Considering proxy servers arre throttled to simply one connection., then farmed out to the rest, why not just "fake" a proxy connection that connects to a dedicated server?

Actually, what I am concerend with is the following. And this is based on my perception as a customer of this game. I preordered this game almost a year before release. The first beta release. I bought 2 copies.

Perceptions:

1. I feel as though the inherit issues of the way this game makes its connections causes more issues to be fixed than client-server.

2. SD tries to fix this game, but (and you can pretty much read thier comments about patches) constantly raise expectations about the fixes/demigods..then lower them by upselling thier new projects and reminding us they have other jobs and that is why fixes are not ready. (I am a customer of this game...so its a bit disrespectful, and what does that tell me if I buy the game they are upselling?)

3. AS it is apparent that the next killer app they are working on will continue to take precedence, what happens when the connectivity services of this game start to phase out? Put the ole cdrom back in the box and trow it out? Or Himachi and find friends?

Final thought: Let the community host thier own servers and let us mod it. The game will die whe we finally let it...free markets work.

No one is saying SD host them, nor am I implying that they can just snap thier fingers, but when you put out a base product that players can host and mod...it grows and then it morphs and there are thousands of programmers out there that can make it grow WAY faster and bigger than a small team. ( I bring Linux as an example again).

Quoting kitch45,

reply 27

Considering proxy servers arre throttled to simply one connection., then farmed out to the rest, why not just "fake" a proxy connection that connects to a dedicated server?

THAT IS MY POINT...I am sorry I could not make it so short and simply as you.

That is an example of many minds thinking through a problem. Efficiency.

Quoting kryo,

reply 26

This is really not the point but ok...let take that. THE CLIENT is getting info from all of its established connection. In this instance they will get thier info from 1 source instead of many if you connect to the "Server" like a proxy. Now. I was not intending on discussing How to program, more of WHY to program. This is about game longevity, increasing customer base and KEEPING them. Read the forums and I think you will find that the connectivity problems, people raging and quitting are the single biggest issue of the game. You can do a few things to address this. "Software Vietnam" meaning...keep pouring resources to fix pieces and keep pouring more time...and effort. B. Leave it as is as customer base tapers off. (Maybe that is the plan: to move us all along because this in not a franchise game) C. ENABLE the community to move it along.

How do you become Blizzard? One day of smart programming at a time.

Quoting Ironrock,

reply 25

...But the idea this is some multimillion dollar project that just is too enormous is pure bs.

If you re-read my post correctly, I set the ballpark figure in the hundreds of thousands, not millions, of dollars. Over-exaggerations aside, the basic idea you're now toying with is already in the game and can be enabled with a mere shortcut console comman, however all this changes is the stability of the connections (assuming the 'host' has the extreme level of bandwidth needed) - as the upload requirements for each 'client' are the same as the Proxy servers minimise the number of connections, however they don't alter the volume of data being sent. 5v5 would most likely still be unplayable on the vast majority of the connections due to the synched nature of the game.

In order to alter this structure, a large scale coding re-write would have to be completed - we're not talking about a mere change from multiple connections to one connection either, we're talking about a massive change in the way the game constructs, sends and processes it's multiplayer information - and as we've covered already, all of this well outside of the realm of possibility.

Quoting ZehDon,

reply 15

Quoting Ironrock,
reply 14
1. Legal ramifications? Enough said on that. And Blizzard did not start Massively, they have evolved to the point they are now through continuing engaement and coding. You clearly do not understand how a business is built.
Firstly, Stardock is the North American Publisher of Demigod, and as thus outside of the North America you're dealing with Atari, I believe. The legal rammifications of such sweeping changes being entirely funded by Stardock are huge. Even a multi-investor change of this magnitude, assuming of course GPG and Atari sign off on the deal which they won't, would essentially place three companies in the midst of a financially irresponsible situation. Atari, being a publically owned company, is under requirement to minimise financial risk and maximise profits; their shareholders wouldn't be too happy with them for spending hundreds of thousands of dollars on a game that has already reached profitable levels of sales and received favourable reception upon it's release. Demigod isn't a gold mine waiting to be tapped, regardless of your feelings on the matter.

Not one sentence here of the business structure implies LEGAL ramifications. That is the investment and profit centers of a business. A programmer making a Dedicated server for users to download and host, encouraging them to get friends to buy the game is not A. SWEEPING change nor is it B. A MASSIVE undertaking. C.

Also...I would appriciate 2 things from you in the future.

1. Please give me an analysis: programming time and business contracts for a modification of the net code for a player to host his own game without going through the matchmaking service of SD and GPGnet...because you think somehow this alters the financial structure of SD, atari and publishers and make it a LEGAL issue (which means you have NO understanding of business. I would so love to know how you get hundreds of thousands of dollars).

2. Please do not explore my feelings as to whether I feel this is an "untapped goldmine" of which I have never said.

If your thoughts were the consensus of the American general population, then we would be a third world country. People make things happen. A programmer with tools and code base can do just about anything. More programmers can make just about anything faster. Once again...and I am so trying to get you to understand this, RELEASE a simple stand alone server component and mod tools and this game will grow. No business will turn down continued revenue. And this is not a BUSINESS REORGANIZATION....and you implying that just means you are not grasping the concept of operation here.

OMG, you are doing it again. That is complete bs. I am getting updates from every player when I play...and so are they. If I get a downloadable server hosting component and purchase time on a game server, then it will most certainly work. How big do you think the packets are???? You are trying to convince me that poor SD cant possibly do this? Ridiculous. Furthermore, you think this equals a business reorg, which it does not.

I am not convinced that somehow Demigod is a special game...none in the world like it. Its packets are SO BIG that none can host it. Its just the way it is. Well, that contention is ridiculous.

And what exactly do you call Extreme level of bandwidth? I am in the business, my friend. I have thousands and thousands of transact-sql statements ensuring the ships are routed into the port of Charleston on time and so on. The company that I own, wrote this program, which the USCGS uses and the Dept. of Homeland Security. Aside from all of the Voip systems we own and host. I think you have a limited understanding of bandwidth and Server side processing. Too many games out there with Too many ded. servers with lots of players working to try to make that point stick.

Once again...making a simple dedicated host program that one can have hosted anywhere does not equal hundreds of thousands of dollars..business reorganizations, LEGAL ramifications or any of the such.

What part of release this to the community do you not comprehend?

Maybe you should recalibrate your reading skills and look at my final note on the original post as opposed to suggesting I am toying with a new idea. Sd can stay just like they are, just give me an option to connect to a dedicated server. Simple.

I don't really have the knowledge to strongly agree or disagree with anyone on how feasible this kind of project would be. What I do know is that GPG is content to have a game with their name on it be unable to provide the basic matchmaking and connectivity functionality which Warcraft 3 provided back in 2002.

however all this changes is the stability of the connections (assuming the 'host' has the extreme level of bandwidth needed)

Lemme also point out that this is pure nonsense. This game has less going on than Starcaft or Warcraft III... DoTA was a Warcraft 3 mod; I was able to play both of those games on dialup without game-based connectivity issues. I literally played Warcraft III on a shared dialup connection. When I was living in a barracks (basically a dorm setup) my roommate and I played simultaneously on 56k without issue.

So when you say that the host needs an extreme amount bandwidth, there's your problem. That should be fixed, and if GPG isn't interested enough in maintaining their image to invest the funds to do so, then I simply won't buy more of their games.

What bugs me the most, and correct me if I'm wrong, is that GPG.net existed before DG's release and during its development. The decision to forego its usage was intentional and this is the result. From what I understand, the publisher was forced to completely rebuild their connectivity, which seems really bizarre to me because I'd always thought the publisher did very minor amounts of actual codework.

Thank you for the point, Obs, you are correct. I feel the same way about what I will purchase in that regard.

One other thing is that back to the intitial idea of people subbing in for players who drop, Left 4 Dead's is pretty much a working model of your proposed setup. That configuration does work really well for that game because L4D's gameplay makes everyone pretty faceless, and also because everyone's health and progress are reset quite frequently.

When you leave a character in Left 4 Dead there's very little impression of you attached to it. You can pick your gun, you can be low on health, but there's nothing to really identify you as the previous owner the way spec, level, or item setup can. Also even L4D tends to stall out at times, with people joining a game, seeing it's hopeless, and then moving on to the next one. I would think that in most cases of disconnect in DG very few people would be inclined to take the reins in a situation which merited abandoning them in the first place. Also rather than have another player fill the spot if it were a simple disconnect, I'd think the player who DC'd would likely be pretty eager to take their spot back as soon as the issue were resolved.

That is a valid point. Having such a component would allow the host and modding community to work on this. If anyone would like to see the modding community in action...simply go to moddb.com and you will see the level of sophistication these folks really have.

Quoting Ironrock,

reply 29

Quoting kitch45, reply 27Considering proxy servers arre throttled to simply one connection., then farmed out to the rest, why not just "fake" a proxy connection that connects to a dedicated server?

THAT IS MY POINT...I am sorry I could not make it so short and simply as you.

That is an example of many minds thinking through a problem. Efficiency.

I don't know much about the mechanics behind it all, but had intended to post a similar idea to the one you posted.

Basically the premise would be as simply as someone has explained to me.

Proxies allow for people to have one path for their bandwidth to go. This is then split as normal to all the players as per usual, so the proxy sort of acts as a fake "player". The bandwidth is the same since you still haveto communicate all the information up and down the pipe with say 5 people, but is only 1 connection instead of 5 potentially flaking ones.

What if everyone was to connect using the "proxy server" kind of connection, but faking the proxy server for a dedicated server?

Could it act in a pseudo client/server capacity?

Could the bandwidth then somehow be minimised with the fewer simultaneous connections?

I may need to draw a crude ms paint drawing to visualise the idea.

The console command I mentioned above activates this exact functionality, the issue is then that the 'server' is a player in the game as there is currently no method for a dedicated server, and the netcode simply doesn't support dedicated servers. The bandwidth requirements are not lessened - only the number of connections - nor is the situation changed from a P2P to a Client/Server as the game is still synched. Thus, players are still able to mess with the LUA files in real-time and create the desync exploit, one person's laggy connection still slows down the game for everyone and as the connections established for Demigod are actually quite decent, the gains of using this type of connection setup are strictly minimal unless you yourself have connection issues.

For this to work in the limited capacity, at the very least, we'd essentially need a 'spectator slot' per match and for the game to support the ability for the spectator to host the game. For this to function is a real world situation, the Lobby would be auto-created and the Spectator slot filled by the server itself. Of course, no options would be able to be changed unless you had access to the server. If we had this, then yes - theoretically from a network perspective - it would be possible. However, as I pointed about above, this merely replicates a Client/Server situation and grants none of the benefits of such a system because Demigod's engine simply isn't built for it and thus can't use it. To grant it such functionalty is - as I've said - an unrealistic outlook because it's actually a massive change to functionality of the netcode within the engine itself.

The 'spectator slot' could also be a game sync place-holder, a point of reference incase of dropped/desynced players. Interesting.

There are new games coming out. I loved Left 4 Dead (1), but it had nowhere near the long appeal that Demigod has. After the first few times through the maps, it's pretty much toast. No LAN in the pc version of Modern Warefare 2 will limit my long term interst. Besides, those are FPS games. The RTS element with FPS and RPG combo in Demigod makes it a winner for me.

IF THIS ONE GETS FIXED, Demigod 2 would be a saleable franchise. Get me a solid, stable, unexploitable game, and you can SELL me expansions, or your next game. It's called return on investment. If there isn't the interest/financing to fix this game, we, the customers have been short changed, and Demigod will be less than it could have been.

There are new games coming out. I loved Left 4 Dead (1), but it had nowhere near the long appeal that Demigod has.

Well the long-term appeal of Left 4 Dead was versus. If you did enjoy versus mode, you should definitely check out L4D2, the improvements are massive. I could write out a laundry list of L4D1 issues that were fixed in the sequel.

Lemme point out one other issue with L4D while we're (I'm) talking about it. Lots of servers ran annoying mods that altered the game mechanics and map setups, it was only by sheer volume of servers that you were able to pretty much ignore those crummy ones. If players hosted games on private servers for a game of this size I do worry that you wouldn't have an acceptable place to just play normal games with standard, established rules.

That may sound kind of irrelevant to you if you personally would set up a good server with the standard ruleset, but that's not something I'd personally be in a position to do, so I'd be apprehensive about it.

Why are we even talking about this?

Why are we even talking to this guy?

Quoting obscenitor,

reply 43

Not to be rude Obscenitor, you're one of the level headed people around here, but it's a non-issue because Demigod is never going to have servers.

Quoting ZehDon,

reply 46

Not to be rude Obscenitor, you're one of the level headed people around here, but it's a non-issue because Demigod is never going to have servers.

Yeah, that's why I was using the subjunctive. :/ I agree it's not something that's ever going to happen, I really think our best bet is hoping that they tie DG into GPG.net in an expansion.

I just find it very interesting that some folks can speak on behalf of what Sd will or wil not do. It is best to speak of what we would like to see happen and the things that are doable to SD (them being the best judge) can be done, or those not, then wont be. I should not be, but am amazed at the constant apologist, and statements as to why nothing can be done about anything. If you are not a memebr of SD, then the comments are pure conjecture as to whether they would release such a component to the community. Any speak of business models, legal ramification, etc is pure conjecture and, quite frankly, a waste of time. And if you think it is a conversation not worth having, then why take the time to click reply in the post? Simply move on to a post in which you might have a valid informed opinion. I also can't help to notice the level of constant negativity behind most people that post. Maybe it is the current culture of younger people or just the culture of the average player in demigod.

Insulting younger people or the community as a whole to explain the negative reaction to a completly unrealistic idea created by someone who clearly has limited understanding of what they're suggesting is kind of grasping at straws, don't you think?

What's honestly more likely:
You've stumbled upon a solution to all of the perceived problems with Demigod's multiplayer component that costs nothing, takes no time to implement and can be done within mere days, and that this solution was somehow missed, overlooked or not even considered by the programming ninja's at Stardock - ninja's who created an international proxy network system for a game not designed to use it and did it without making a single change to the game's coding and did it within two weeks that resolved one of the worst multiplayer centric title launches in recent living memory and that no one will listen to your sheer brilliance because they're both younger than you or completely negative minded and refuse to give any thought to what will be considered the idea that 'saved' Demigod and allowed to go on to prosper and eventually be worth millions of dollars in sales to it's respective owners
or
We're saying that the idea is simply too unrealistic a demand and wouldn't work because the idea is simply too unrealistic a demand and wouldn't work.

I've said my part, I won't reply further.
Best of luck.

Ironrock you are aware that the reason this game is peer-to-peer is because it was made using the supcom engine wich was a simulated game, as is demigod. As far as i know gpg has no games that use client-server connections and have no reson to invest in that modle as would not benifit there other games.

Because supcom was simulated and due to bandwith limitaion the games needed to be syncronised as live updates of the position of every unit would have been beyond any connection, though demigod may not need this the investment to change it would be massive and see negative impacts on connection prefomance. I am guessing that networks do not have a bandwith limitaion so low and therfore you underestimating the problem increasing bandwith requirements requirements outside the US.

The Forums Are Now Closed!

The content will remain as a historical reference, thank you.

How to fix the Multiplayer in Demigod