Howdy, Stranger!
It looks like you're new here. If you want to get involved, click one of these buttons!
Quick Links
Cease The Lag Debate
Andromedus1
Member Posts: 67
I understand that many are skeptical of Darkfall's claims that it can handle 10,000 simultaneous players and battles involving hundreds. They base these opinions on current games and current technologies, and the current level of care (or lack thereof) that the giant corporate MMORPG market has fostered in its community of programmers. Please take the time to read the following quotes, taken from the following links: http://darkfallinfo.com/index.php?page=Info&code=darnvcusyc and http://darkfallinfo.com/index.php?page=Info&code=u1gtj5gd6g If you choose to call these men liars, then it is obvious that nothing will satisfy you until at least open beta is out, to which this community must ask: what are you doing here? If you have the patience to write endless garbage about a game you don't want to play, then you ought to have the patience to read the following page or two worth of explanation.
First Interview
MMORPG.com: Your game boasts the ability to interact with tens of thousands of players simultaneously. Technically, how do you plan to achieve this?
Tasos Flambouras: The decision that Darkfall would support a high number of players came early in the technical design process. It was obvious that the numbers we wanted would be hard to support on the server end without clustering, so we started evaluating available clustering technologies. None of them met our requirements at the time, and we soon understood that we would need to do this ourselves. After a lot of brainstorming and experimentation, we arrived at a model for distributed computing which we’ve pretty much stayed with since then.
Throughout every part of game development, we’ve designed and programmed using asynchronous communication in between very small, movable modules. This has imposed restrictions on the programmers, but it has also helped error location and debugging immensely. We can spread the game servers across anything in between one and several thousand computers. Additionally, we can move modules in between computers while the game is up and running, making load balancing a simple matter of looking at resource utilization before moving modules around.
Another problem with a high number of players is the limited resources of player PCs, and limitations in internet connectivity. There areseveral levels of optimizations intended to reduce lag and problems due to these issues. Among these is what we like to consider a quite innovative mechanism for dynamic level-of-detail adjustment, of course in addition to both context-aware and plain binary compression mechanisms for network data.
The combination of a distributed model for servers, which will allow us to always have enough CPU power, and inventive optimizations that can balance CPU usage against network performance, gives us a mechanism where resource limitations could be solved by adding a few more servers.
Second Interview
MMORPGDotCom: Darkfall will probably be one of the first so-called “ultra mmorpg's”, allowing for about ten thousands of players to be online in a single persistent world at once. At the same time you promise us player collision (you can't walk through other players), an extremely large world with high-quality graphics and world-physics. Without getting too technical, why do you think your team will accomplish all of this, while we see that current mmorpg's like Shadowbane can barely handle the lag already? What's the team's experience with these types of programming?
Claus Grovdal: We don’t really want to comment on what other games are doing or not doing, but in layman’s terms what we have done is that we first built the system that makes this possible - from scratch, and then built the game on top of that. It usually happens the other way around.
It’s impossible to explain this further without getting a bit technical, so here’s a more detailed explanation by Erik Johansen, our network programmer:
Our server and networking technology, which we've dubbed a "Real-Time World Simulator" or RTWS for short, has been built from scratch using a design that's in between the peer-to-peer way of thought, and the newer Grid computing concepts. What the users will see as a single server, will in reality be a series of many similar computers, dynamically and continuously moving tasks around in order to keep the workload evenly distributed and timely processed.
We are able to simply add a new preconfigured computer to such a server cluster, and have it join its capacity to the already running world simulation without interruptions for the users. And the other way around, if one of the server computers fails for any reason, the remaining ones will take its workload, distribute it and continue after some seconds of reorganizing.
This scalability will be an important asset for us when we work to keep servers as responsive as possible for the users. In addition we do of course have prediction algorithms on multiple levels, from simple dead reckoning to advanced analysis of player behavior. The rest is merely about having enough bandwidth to keep up with the network traffic.
Some of Erik’s related experience over the last 10 years includes developing an internal content distribution system for a large industrial company serving thousands of simultaneous users, participation in the development of networked communications software used daily by more than 100K users, and development of software for distributing specific tasks among several nodes within a severely reduced/damaged network, with a focus on redundancy and failover mechanisms.
Comments
i'm sorry but 10,000 at once is just plain rediculous. i dont care how friggen good your programming is there is far too many variables that interfere with the game to make a lag free game with that many players. Battles with hundreds? thats achievable but some cuts needs to be made (graphics and effects mainly).
you can argue all you want but until we get decent connections, better servers and even better programming we wont see lag free games.
good luck to Darkfall and i can see maybe up to 100 at once (damn big fight) could be lag free (with those graphics and sparse landscape)
MMO wish list:
-Changeable worlds
-Solid non level based game
-Sharks with lasers attached to their heads
You're telling me that in six short minutes, you spotted my thread, read my entire post, analyzed what it had to say, decided you were gainst it, and then composed and submitted your post...
Next time plan a little more carefully and you won't look so ridiculous. It's safe to say your credibility just went out the window, maybe you can follow it to some other forum?
How about, I'll believe it when I see it in action. Only thing I have faith in is God, Aventurine's going to have to prove to me that this works.
"True friends stab you in the front." | Oscar Wilde
"I need to finish" - Christian Wolff: The Accountant
Just trying to live long enough to play a new, released MMORPG, playing New Worlds atm
Fools find no pleasure in understanding but delight in airing their own opinions. Pvbs 18:2, NIV
Don't just play games, inhabit virtual worlds™
"This is the most intelligent, well qualified and articulate response to a post I have ever seen on these forums. It's a shame most people here won't have the attention span to read past the second line." - Anon
Innocence is so cute. Life will take care of erasing it.
Why not cease any debate about DF then? Obviously, we should all believe whatever is being told to us by Aventurine, since obviously they would never tell us anything but the truth, right? I mean seriously. These guys are in no way interested in selling their game. They are just a bunch of cool dudes who work tirelessly day in and out, having no other goal in mind than to please US, regardless of the cost, work, sweat or blood involved.
Come on. Can't you all see it? There's dozens of companies out there, all money-grubbing, corporate sell-outs. Aventurine is the shining beacon of hope, the dream of MMO-Players the world around and there is NO WAY IN HELL that they could POSSIBLY, be saying ANYTHING but the full truth in all their announcements, updates, and blogs.
P.S.: On a serious note, it's recently become a trend to put up developer interviews and then point to them as proof for features in Darkfall. Dear fans, an interview is not proof of existance...it is proof of intent. I have no doubt Aventurine intends for their technology to work the way they say, but will it? Only time will tell.
Which Final Fantasy Character Are You?
Final Fantasy 7
The absence of evidence is not the evidence of absence. oO
Can Aventurine do all they claim? Yep they could. The possibility is there otherwise none of us would even be in this forum.
Will Aventurine do all they claim? No one knows. We "All" have to wait till launch to find out.
The absence of evidence is not the evidence of absence. oO
Can Aventurine do all they claim? Yep they could. The possibility is there otherwise none of us would even be in this forum.
Will Aventurine do all they claim? No one knows. We "All" have to wait till launch to find out.
Exactly my point. But on this forum, DF is either treated as a game that's already delivered all it's promised (with "evidence" in the form of developer interviews and blogs) or, on the other hand (and just as ridiculous), as Vaporware.
I'm not saying DF can't do it. I'm saying DF hasn't done it -yet-. So telling people to cease discussing possible problems is very premature.
Which Final Fantasy Character Are You?
Final Fantasy 7
grid style computing variable it opens up abig can of worms. Doing something like this means a couple of things as far as i know:
Increased latency for everyone since each server portion of the grid will have to update and wait for the response from the other one. If one has a delay they all will suffer from it.
Scalability issues: No software is perfect so while trying to take advantage of real time grid computing some of the servers hardware power will be lost.
i duno im tired i just got off work but hopefully the DF figured it out.
Look into the guy responsible for the network coding for this game and his previous experience / credentials.
(I have already provided them in other threads)
They make these claims extremely dubious in my opinion, but I’m willing to wait and see if they come to fruition.
Oh yeah and the “Cease the discussion” because Tasos told us everything is going to be perfect makes you sound like your in a cult and you’ve been brainwashed by its leader in my opinion.
And even worse it was not always 10,000 users , years ago in other interviews it was a max of 50,000 -60,000 and they where planning for 30 to 40k.
I’m sure the current 10k number will also again be significantly reduced during beta.
Can you give me a quote/link on the numbers you mention?
Here is a interview from Razorwax time in 2001 when they have 10000 players online at same time as their aim. So the goal 2001 was 10000 which seems to be the same as now.
Now your putting numbers of 30-60K online on one server same time in the developers mouths, so a link to it would be appreciated.
www.darkfallinfo.com/index.php?page=Info&code=d6yj4ybca7&highlight=interview
Q: How many players will the gameworld support?
A: We have no idea at this point, and guessing is sort of pointless. But hey, Ill guess anyway. We are aiming at 10.000 on each server cluster, hoping for 20.000 and dreaming about 30.000 players at once.
http://darkfallinfo.com/index.php?page=Info&code=ydyy54gmwt
Jonric: Is it your intention to have one large gameworld or multiple smaller ones? Do you have a target server capacity?
Erik Sperling Johansen: For starters we're going for one large server cluster, with all users living in the same gameworld. The design of the server permits scaling to an unlimited size, but we will meet practical hardware limits giving us a theoretical max of about 50-60 thousand simultaneous users. Until we get this properly tested that number is just an educated guess though, if the server can handle 30k users, I'm satisfied, if it can handle 40k I'll be happy.
Remember this is apparently going to be a huge seamless world (except for dungeons) that is “bigger then any other MMO” full of custom created content that has FSP real-time type combat running physics based features like weather and ballistics etc. with advanced NPC AI / path finding, dynamic MOB spawns and city building/destruction, resource gathering, questing, crafting, naval battles that will achieve excellent performance and FPS during battles of 100-200 people with only “average“ machines, all on a server that is hosting at lease least 10,000 users concurrent users adequately.
And please don't mention Eve is not seamless its separated into lots of separate nodes or zones, which have server/network resources dyanamicaly assigned to each one depending on load, totaly different thing.
This sounds totally impossible to me, and I’ve never seen it done before, but like I said I’m willing to wait and see.
And based on what I think I know about Erik Sperling Johansen previous experience and credentials I find it very hard to believe he will be able to achieve this, but like I said I’m not saying he definitely wont, but I will only believe when I see it.
My prediction is that this 10,000 number will be significantly reduced again from their previously even more ambitious numbers to something much more realistic after they actually attempt to get thousands of real players from all over the world attempting to play test, or even before that during the hardware test.
Thx for the link.
Seems they settled for 10000.
What you have to remember is that the game they is talking about in those 2001 interviews isnt the same game that is in development 2008.
Darkfall 2008 have very little to do with the old Razorwax game. The gameworld is totally different, graphic engine is a totally new one.
So interviews from Razorwax time isnt accountable as i see it.
Eh, notice how the quotes are all in past tense? As in, the server technology has already been built, then the game was built on top of it? Er wait no, they possibly could just be lying about the server technology that they had to build a game on top of, since the game's not real either.... wow.
Vanguard and AOC really shoved their member up some people's pooper and left some long lasting scars huh?
Eh, notice how the quotes are all in past tense? As in, the server technology has already been built, then the game was built on top of it? Er wait no, they possibly could just be lying about the server technology that they had to build a game on top of, since the game's not real either.... wow.
Vanguard and AOC really shoved their member up some people's pooper and left some long lasting scars huh?
Yeah yeah. Sure. *smirk* Someone says something you don't like, you label them a Vaporware troll, eh? Notice how I never said the game isn't real? I've never said that so far and I'm not gonna start now, even if some of you die-hard Fanbois want to just dismiss any possible flaw/problem of the game.
Fact is, you don't know if they really have a game that can handle the amount of players they say it can, at the performance they promise. Or can you? No? Thought so.
Which Final Fantasy Character Are You?
Final Fantasy 7
I thought it worth mentioning that EvE online's single server world can happily handle 25 - 30 Thousand people online at once.
So yea it is possable to run with 10,000 but it needs a scalable server with decent programming and the abilaty to add extra hardware when needed.
Another great example of Moore's Law. Give people access to that much space (developers and users alike) and they'll find uses for it that you can never imagine. "640K ought to be enough for anybody" - Bill Gates 1981
Then why did you use a link to an interview from 2001 from the time of Razorwax for the source of 10,000 concurrent user number?
Anyway I hope they can achieve 10,000 but I very much doubt it, as far as I’m concerned nothing the developers have said is certain until we start playing Darkfall.
Eve is not a huge seamless world, its made up of lots of connected zones or nodes, not the same thing.
Hmm... I'm confused.
When you doubted Polarization's numbers you used a quote from RazorWax circa 2001 to prove him wrong. Yet when he provided a link from that same time period in an interview with a Dev.; then that time-period is no longer valid? Then why did you use it as a counter-point?
Re-reading Polarization's original post... he clearly says:
"...years ago in other interviews it was a max of 50,000 -60,000 and they where planning for 30 to 40k".
His statement was indeed true as was evidenced by the link he provided. He was using an example of: just because something is said once doesn't mean it isn't subject to change. The numbers were once targeted to be 30k+. Now they are targeted to be 10k.
He is using that example to show that he believes (as in his opinion) that 10k number will be further down-graded before the game goes live. Considering the reasoning behind the opinion and what we know of current MMOs... that is not a far-fetched opinion.
On the flip-side...
There is the possibility that Aventurine might advance the server technology to allow lag-free play with the numbers they are currently aiming for. That is also a valid opinion as it is what the developers are saying.
Which one will be closer to reality? We'll have to wait and see what happens.
didn't Wish achieve this before it ran out of funding and flipped and flopped.
Doesn't WWII online also boast this same ability with just as complicated combat mechanics if not more as DF is boasting? and that was created when 2001? i remember reading they had some problems but they were not problems impossible to overcome.
If you look at the information I have found about the Developer who is responsible for this server technology I personally don’t think it is a very valid opinion, but of course it is a possibility that he “might advance the server technology” single handedly but I doubt it myself.
So I will wait and see instead of believing eveything the developers say without question.
Especially when they have already revised the number of concurrent users at least once already, so I see no reason they may not do so again, especially during open beta.
The absence of evidence is not the evidence of absence. oO
Can Aventurine do all they claim? Yep they could. The possibility is there otherwise none of us would even be in this forum.
Will Aventurine do all they claim? No one knows. We "All" have to wait till launch to find out.
Exactly my point. But on this forum, DF is either treated as a game that's already delivered all it's promised (with "evidence" in the form of developer interviews and blogs) or, on the other hand (and just as ridiculous), as Vaporware.
I'm not saying DF can't do it. I'm saying DF hasn't done it -yet-. So telling people to cease discussing possible problems is very premature.
Only people who treat it that way are overly talkative noobs who get made fun of by everyone else
Clueless mate that is what you are, the technology exists already and has existed for years. With the built in scale down of graphics for large battles it i more than acheivable.
________________________________________________________
Sorcery must persist, the future is the Citadel
I was in Wish beta but I don’t remember anything about them achieving 10,000 concurrent users and that game was canceled after 2 weeks of beta, do you have source for this?
And wish did not have real time FPS type combat and many other features that DF has so its not a good comparison.
WWII online has always suffered from performance and network, latency and prediction code problems.
It also has a hardcoded maximum number of visible players well under the 100’s Tasos is talking about.
The number one player complaint for WW2 online apart from side balance has always been performance performance performance, the client and game is extremely taxing on system resources and causes most people to play with under 30FPS in “large battles” even with powerfull computers
It also suffers from lag problems that are inherent with all games with client side hit detection when ping times start going over 100 ms.
It also has very low level poly and textured models and generally very Spartan graphics to achieve this.
I played WW2 online for 5 years from release, they have achieved a great deal but its far from perfect.
DF with WW2 Online’s network code and servers, would fail miserably in my opinion, it just would not work becuase they are different types of games.
DF has to do so much better to achieve anything more then the sort of subscriber base WW2 online currently has.
Eve is not a huge seamless world, its made up of lots of connected zones or nodes, not the same thing.
They are or will be using the same mechanic AC1 used which is calculated on an horizon, zones into dungeons. But the game world will consist of a large cluster of servers.
________________________________________________________
Sorcery must persist, the future is the Citadel
I'm willing to give people a pass that want to believe what the developers are saying. Until there is proof that they are lying... it is still a valid opinion. I agree though that healthy skepticism isn't a bad thing.
The phrase from the original OP's post that worries me the most is:
"Tasos Flambouras: The decision that Darkfall would support a high number of players came early in the technical design process. It was obvious that the numbers we wanted would be hard to support on the server end without clustering, so we started evaluating available clustering technologies. None of them met our requirements at the time, and we soon understood that we would need to do this ourselves. After a lot of brainstorming and experimentation, we arrived at a model for distributed computing which we’ve pretty much stayed with since then."
Unless I'm reading that wrong... that says to me that a group of devs working on an MMO developed a better clustering technology than ANY other techology out at the time... or developed since.
Considering the clustering technologies made and developed by IBM, SUN, HP, etc... I am a bit skeptical of that statement. The one word that I might be misunderstanding in that phrase is "available".
Perhaps clustering technology developed by those other companies was not "available"... means it was not affordable, or something like that. So, if "available" means that... then I could see them wanting to develop their own technology as opposed to being tied to some licensing deal.