Well in congested areas my computer is probably the bottleneck, so it's hard to say. But also consider that LOTRO is not as much of a real time combat game as conan and COS. I can image lag having a greater effect on gameplay in AOC and COS because of their combat systems. Yes, if lag occurs, real time combat suffers alot more then auto-attack combat. As I said - I dont lag in Lotro, why would I in Spellborn? I agree that companies should be forthright with information regarding instancing, but also consider that COS isn't really close to releasing and they may not have even made a final decision about it. Anyway, I understand that you don't like instancing period, but to act like there's no difference between a game like AOC where multiple instances are the norm and another game that may just use it in cases of extreme congestion seems irrational to me. We dont know how large they are gonna make the instances, it might be 50, it might be 400 - we dont know. So making an comparison with a game who use the exact same inatance system, seems rational to me. As I said in my OP - Im only worried. This isnt the end of spellborn for me...
The biggest issue with instancing is that it's relatively pointless since there are algorithms and programs that can dynamically add servers to clusters for the same 'instance' of a given thread and/or program/process. A zone in this case can be treated similarly, so why don't they do it? It seems like a waste of time to make instances that are mirrored to other clusters when you just make a 'pool' of the servers instead that you can queue up any old time with a "pool limit." Pardon me for making it geeky, it's just something I find odd when I'm looking at my Operating Systems text book that has similar issues in regards to the race condition and semaphores for multi-threaded programs.
The biggest issue with instancing is that it's relatively pointless since there are algorithms and programs that can dynamically add servers to clusters for the same 'instance' of a given thread and/or program/process. A zone in this case can be treated similarly, so why don't they do it? It seems like a waste of time to make instances that are mirrored to other clusters when you just make a 'pool' of the servers instead that you can queue up any old time with a "pool limit." Pardon me for making it geeky, it's just something I find odd when I'm looking at my Operating Systems text book that has similar issues in regards to the race condition and semaphores for multi-threaded programs.
-- Brede
I have been up almost 40 hours (thank you insomnia...) and then I read this.
I think you caused my one last brain cell to die from the stress of making me think...
There are 3 types of people in the world. 1.) Those who make things happen 2.) Those who watch things happen 3.) And those who wonder "What the %#*& just happened?!"
Well in congested areas my computer is probably the bottleneck, so it's hard to say. But also consider that LOTRO is not as much of a real time combat game as conan and COS. I can image lag having a greater effect on gameplay in AOC and COS because of their combat systems. Yes, if lag occurs, real time combat suffers alot more then auto-attack combat. As I said - I dont lag in Lotro, why would I in Spellborn? Sorry to interviene, but, the reason is due to the fundamental design of the game. Basically LotR, WoW and many rpg without the need for real time combat use a predictive code to emulate in advance where the characters will be, the clients do it individually to give a smooth feel and the server will do its own, the servers decision is final but WILL differ from that of the clients. In LotR etc this will have little effect on gameplay so long as when you attack the object is 'in range' from the servers perspective (this is why you will sometimes see your fireballs shooting massive distances, unhittable players on your screen right in front of you, etc). For games such as TCoS, AoC and other such combat this can not be done. The cade can not be allowed to predict the current position of the characters, purely for asthetic reasons. Some games get around this by using a form of client side detection to take the stress from the servers. This however opens the game massively to hacking and imo is a no go from the begining. As it stands currently the hardware doesn't fit the players expectations and instancing zones (to limit the number of players present) is a lesser of 2 evils. However I'm with the guy above who remarked that 400 in 1 instance at the same time with real time combat, believe it when I see it. Personally I dont mind the instancing but I can see how others do, it could be handled better than it currently is in AoC to make it more seemless and less immersion breaking but they have chosen by design to go the other route and make it totally open and visible to their players. I feel its like comparing zoning from WoW to EQ2. In EQ2 you see it all from one zone to the next you see the load, in WoW they hide it. The same sort of thing could be done with instancing but nobody is. They need to fool the players more imo. I agree that companies should be forthright with information regarding instancing, but also consider that COS isn't really close to releasing and they may not have even made a final decision about it. Anyway, I understand that you don't like instancing period, but to act like there's no difference between a game like AOC where multiple instances are the norm and another game that may just use it in cases of extreme congestion seems irrational to me. We dont know how large they are gonna make the instances, it might be 50, it might be 400 - we dont know. So making an comparison with a game who use the exact same inatance system, seems rational to me. As I said in my OP - Im only worried. This isnt the end of spellborn for me... On this I agree, without clientside hit detection I find i highly unlikely they will have 400 chars in a single area, but then I would love for them to prove me wrong.
----- The person who is certain, and who claims divine warrant for his certainty, belongs now to the infancy of our species.
The biggest issue with instancing is that it's relatively pointless since there are algorithms and programs that can dynamically add servers to clusters for the same 'instance' of a given thread and/or program/process. A zone in this case can be treated similarly, so why don't they do it? It seems like a waste of time to make instances that are mirrored to other clusters when you just make a 'pool' of the servers instead that you can queue up any old time with a "pool limit." Pardon me for making it geeky, it's just something I find odd when I'm looking at my Operating Systems text book that has similar issues in regards to the race condition and semaphores for multi-threaded programs.
-- Brede
Some game engines do, some are better than others (and their price tags are also). However there is always a limit no matter how many servers are added to a cluster. The easiest way to explain this is as always through referal of extremes.
If we take for expample there are now 400 players in a zone and our game is so advanced our cluster management algorithm has decided that each player in this zone requires its own seperate server, thats 1 server per player (something that few companies will be doing ). Now aside from needing a server room the size of a small country and the resources of its gdp to maintain, the benifits from doing so are reduced each time a new server is added. While its true that the extra processing power will help greatly to alleviate the actual processing power required, the extra traffic and noise is being increased exponentially. Remember all the time the server has checks and balances (to put it stupendously mildly) that have to be run each and everytime anyone so much as moves, basically it has to continue to run the game. An mmo is basically servers creating a world that clients then play on, npc's quests etc (different games do it different way to reduce server load, each though at a price).
Now lets take the other side. An fps works a very different way (most anyway). The clients create the world which the servers then act as a translator/link to put other clients in that world, each one independantly. This is a far more efficient way of doing things. But remember there are no quests, no npc's, no mobs. The clients are more than capable of generating these, its not a problem it does so in all single player rpg's with far superior game play, graphics, mob ai etc but each set of world rules will be stored client side. This means everyone can make up their own game world rules. Want a few hundred plat, want to wipe out that entire dungeon, want to get that mount? simply alter this mem entry, etc.
MMO's therefor need to have the world based serverside, we all agree on that, there is no magical answer from that point onwards. The answer is higher processing speeds for both client and servers and better networking between them. It will come in time but until then we'll have to make do with zones and instances. How companies implement them into the game seemlessly with superior coding I think derives at how well it will do (amongst other other things such as gameplay ).
----- The person who is certain, and who claims divine warrant for his certainty, belongs now to the infancy of our species.
Originally posted by mrw0lf Originally posted by ladyattis The biggest issue with instancing is that it's relatively pointless since there are algorithms and programs that can dynamically add servers to clusters for the same 'instance' of a given thread and/or program/process. A zone in this case can be treated similarly, so why don't they do it? It seems like a waste of time to make instances that are mirrored to other clusters when you just make a 'pool' of the servers instead that you can queue up any old time with a "pool limit." Pardon me for making it geeky, it's just something I find odd when I'm looking at my Operating Systems text book that has similar issues in regards to the race condition and semaphores for multi-threaded programs. -- Brede
Some game engines do, some are better than others (and their price tags are also). However there is always a limit no matter how many servers are added to a cluster. The easiest way to explain this is as always through referal of extremes. If we take for expample there are now 400 players in a zone and our game is so advanced our cluster management algorithm has decided that each player in this zone requires its own seperate server, thats 1 server per player (something that few companies will be doing ). Now aside from needing a server room the size of a small country and the resources of its gdp to maintain, the benifits from doing so are reduced each time a new server is added. While its true that the extra processing power will help greatly to alleviate the actual processing power required, the extra traffic and noise is being increased exponentially. Remember all the time the server has checks and balances (to put it stupendously mildly) that have to be run each and everytime anyone so much as moves, basically it has to continue to run the game. An mmo is basically servers creating a world that clients then play on, npc's quests etc (different games do it different way to reduce server load, each though at a price). Now lets take the other side. An fps works a very different way (most anyway). The clients create the world which the servers then act as a translator/link to put other clients in that world, each one independantly. This is a far more efficient way of doing things. But remember there are no quests, no npc's, no mobs. The clients are more than capable of generating these, its not a problem it does so in all single player rpg's with far superior game play, graphics, mob ai etc but each set of world rules will be stored client side. This means everyone can make up their own game world rules. Want a few hundred plat, want to wipe out that entire dungeon, want to get that mount? simply alter this mem entry, etc. MMO's therefor need to have the world based serverside, we all agree on that, there is no magical answer from that point onwards. The answer is higher processing speeds for both client and servers and better networking between them. It will come in time but until then we'll have to make do with zones and instances. How companies implement them into the game seemlessly with superior coding I think derives at how well it will do (amongst other other things such as gameplay ).
I think the better solution is to move away from the client/server model and use what's used in the G2 and G1 (p2p networks) in regards to making every client a node in the network for processing. It may seem impossible today, but that's what they said about p2p networks as well.
I hate loading screens and zoning. It's a shame this game will be like EQ2/AoC both of which I detest because of their over use of instances and zoning.
I hate loading screens and zoning. It's a shame this game will be like EQ2/AoC both of which I detest because of their over use of instances and zoning.
If you think EQ2 overuses instancing and zoning, you should stay away from Spellborn and miss all the fun.
Originally posted by Delanor Originally posted by Anti-Fanboi I hate loading screens and zoning. It's a shame this game will be like EQ2/AoC both of which I detest because of their over use of instances and zoning.
If you think EQ2 overuses instancing and zoning, you should stay away from Spellborn and miss all the fun.
If that's the case I have no problem doing so seeing as cities like Freeport and Qeynos don't feel like cities at all.
If you think EQ2 overuses instancing and zoning, you should stay away from Spellborn and miss all the fun.
If that's the case I have no problem doing so seeing as cities like Freeport and Qeynos don't feel like cities at all.
Cities like Qeynos and Freeport indeed do not feel like cities at all, but I do not think that is due to the zoning in EQ2. It is due to the zones and the npc's in them being static. Nobody seems to live in those towns. Most npc's just stand there and don't seem to have a life. But I am going off-topic here.
Comments
The biggest issue with instancing is that it's relatively pointless since there are algorithms and programs that can dynamically add servers to clusters for the same 'instance' of a given thread and/or program/process. A zone in this case can be treated similarly, so why don't they do it? It seems like a waste of time to make instances that are mirrored to other clusters when you just make a 'pool' of the servers instead that you can queue up any old time with a "pool limit." Pardon me for making it geeky, it's just something I find odd when I'm looking at my Operating Systems text book that has similar issues in regards to the race condition and semaphores for multi-threaded programs.
-- Brede
I have been up almost 40 hours (thank you insomnia...) and then I read this.
I think you caused my one last brain cell to die from the stress of making me think...
There are 3 types of people in the world.
1.) Those who make things happen
2.) Those who watch things happen
3.) And those who wonder "What the %#*& just happened?!"
-----
The person who is certain, and who claims divine warrant for his certainty, belongs now to the infancy of our species.
Some game engines do, some are better than others (and their price tags are also). However there is always a limit no matter how many servers are added to a cluster. The easiest way to explain this is as always through referal of extremes.
If we take for expample there are now 400 players in a zone and our game is so advanced our cluster management algorithm has decided that each player in this zone requires its own seperate server, thats 1 server per player (something that few companies will be doing
). Now aside from needing a server room the size of a small country and the resources of its gdp to maintain, the benifits from doing so are reduced each time a new server is added. While its true that the extra processing power will help greatly to alleviate the actual processing power required, the extra traffic and noise is being increased exponentially. Remember all the time the server has checks and balances (to put it stupendously mildly) that have to be run each and everytime anyone so much as moves, basically it has to continue to run the game. An mmo is basically servers creating a world that clients then play on, npc's quests etc (different games do it different way to reduce server load, each though at a price).
Now lets take the other side. An fps works a very different way (most anyway). The clients create the world which the servers then act as a translator/link to put other clients in that world, each one independantly. This is a far more efficient way of doing things. But remember there are no quests, no npc's, no mobs. The clients are more than capable of generating these, its not a problem it does so in all single player rpg's with far superior game play, graphics, mob ai etc but each set of world rules will be stored client side. This means everyone can make up their own game world rules. Want a few hundred plat, want to wipe out that entire dungeon, want to get that mount? simply alter this mem entry, etc.
MMO's therefor need to have the world based serverside, we all agree on that, there is no magical answer from that point onwards. The answer is higher processing speeds for both client and servers and better networking between them. It will come in time but until then we'll have to make do with zones and instances. How companies implement them into the game seemlessly with superior coding I think derives at how well it will do (amongst other other things such as gameplay
).
-----
The person who is certain, and who claims divine warrant for his certainty, belongs now to the infancy of our species.
If we take for expample there are now 400 players in a zone and our game is so advanced our cluster management algorithm has decided that each player in this zone requires its own seperate server, thats 1 server per player (something that few companies will be doing
Now lets take the other side. An fps works a very different way (most anyway). The clients create the world which the servers then act as a translator/link to put other clients in that world, each one independantly. This is a far more efficient way of doing things. But remember there are no quests, no npc's, no mobs. The clients are more than capable of generating these, its not a problem it does so in all single player rpg's with far superior game play, graphics, mob ai etc but each set of world rules will be stored client side. This means everyone can make up their own game world rules. Want a few hundred plat, want to wipe out that entire dungeon, want to get that mount? simply alter this mem entry, etc.
MMO's therefor need to have the world based serverside, we all agree on that, there is no magical answer from that point onwards. The answer is higher processing speeds for both client and servers and better networking between them. It will come in time but until then we'll have to make do with zones and instances. How companies implement them into the game seemlessly with superior coding I think derives at how well it will do (amongst other other things such as gameplay
I think the better solution is to move away from the client/server model and use what's used in the G2 and G1 (p2p networks) in regards to making every client a node in the network for processing. It may seem impossible today, but that's what they said about p2p networks as well.
-- Brede
You guys are losing me...I hate working with all that mumbo jumbo...thats why I make things pretty
I hate loading screens and zoning. It's a shame this game will be like EQ2/AoC both of which I detest because of their over use of instances and zoning.
If you think EQ2 overuses instancing and zoning, you should stay away from Spellborn and miss all the fun.
--
Delanor
If you think EQ2 overuses instancing and zoning, you should stay away from Spellborn and miss all the fun.
If that's the case I have no problem doing so seeing as cities like Freeport and Qeynos don't feel like cities at all.
If that's the case I have no problem doing so seeing as cities like Freeport and Qeynos don't feel like cities at all.
Cities like Qeynos and Freeport indeed do not feel like cities at all, but I do not think that is due to the zoning in EQ2. It is due to the zones and the npc's in them being static. Nobody seems to live in those towns. Most npc's just stand there and don't seem to have a life. But I am going off-topic here.
--
Delanor