![:man_facepalming:](https://d38bqls1q93fod.cloudfront.net/en/wow/images/emoji/twitter/man_facepalming.png?v=6" ":man_facepalming:")
I’m going to leave this one here. Player count and spell complexity are separate axis, but I don’t want to go into an algorithm lecture.
Key point is the BLIZZARD said they can do it.
Nope, they said it’s their “goal” and they are “looking into it”. Player count and spell complexity all relate to server load and there is a point they just can’t handle it. 80 people casting complex spells isn’t that far from 400 people casting simple spells.
80 x 25 = 2000^2 = 4,000,000
400 x 5 = 2000^2 = 4,000,000
Yet reasonable feedback is considered by many to be noise.
We can trust Blizzard, of course. :-^)
4 Likes
Mind you I have largely checked out on this thread, but I could not help but read your example Voo.
A so-called level 1 spell will have the same effect server-side as a level 5.
Particle effects are calculated purely on the client-side, hence the studdering and hickuping.
It is true the server has to register and keep track of when any spell is triggered in order to calculate its outcome, and when many are in a small zone can start to affect the server load. The only thing that can add or subtract lag is the compilation of what the spell does on any given die roll.
Now, let’s break down the 400 vs 80. The only difference is the number of threads that are fired off for the computation. They do not equal the same load per say.
It’s rolling against more people as more people are rolling against it.
People say layering is the least bad option, but quite a few people have brought up the clustered subserver idea where you are locked to your subserver, the name pool is shared across all subservers and AH is shared, and the subservers are then merged as the population lowers and stabilizes. This sounds MUCH better than layering and the TRUE least bad option, but most seem to completely ignore it.
2 Likes
Agreed.
My concern is that are Blizzard is not even testing that many people on one layer (or server or shard or whatever you want to call it.)
I did watch Asmongold bring a server to a halt in the last beta stress test. Basically he had everyone line up from SW to Blasted Lands to kite a mob, then they collapsed to SW to follow the Blasted Lands.
I don’t know how many players were involved or how they got them on 1 layer, but… the game was lagging to about 1 frame per 5 to 10 seconds and basically broken. I’m surprised more weren’t kicked. I heard Blizzard was spawning monsters on other beta test realm, or maybe even that one, but you couldn’t see anything on that server with the lag.
So all this testing about layers was a waste if the time could have been spent just to get good gameplay for 100s of players in one zone at once.
I scrolled up to try to figure out what you are saying here.
You clearly do not understand how game servers work.
There is no “complex” spell as far as the server is concerned. Everything is a stream of actions that trigger a method or function that calculates a result that is then fed back to you and everyone else’s clients.
It is true that the more bodies you cram into a computational shard (watch everyone misconstrue the use of the word shard) will cause more load on the CPU. However, having fewer people casting spells does not necessarily equal the same load.
I also believe its you that has stated they simply can’t handle large numbers or some nonsense. The new server architecture is built upon a redundant cloud infrastructure that will spin resources up and down depending on what is needed.
While you might see a hitch or bogging down at the beginning of a large influx of players, as the server (which is not one physical box) begins to spin up more resources and load balance it will go away.
Not ignore it. The thing is this is but many of MANY threads on this topic. We have had some very good discussion around this very topic.
While it’s not a bad solution it also has built-in problems. Mainly not being able to join your friends or guild if you find yourself on the wrong layer. And that is anti-community for sure.
But it has been discussed and its not a terrible plan.
There are trade-offs, for sure. With locked subservers, the trade off is ease of playing with friends and guildies. However, it completely mitigates the various ways layering can be abused for various purposes because you are locked to your subserver/layer with no way to switch period, let alone on the fly. One possible concession would be to allow players to choose their preferred subserver upon first login and then lock them there after that point.
Ultimately, I want the same the same thing the pro-layering crowd likely wants. I want Classic to succeed as the most authentic recreation of the 2006 1.12 Vanilla experience as possible. I just wish there was a bit more consistency in what the Classic team is going for. if the goal to recreate the content schedule, then things like x-realm BGs and the like should be added, but in later phases. Or if the goal was to recreate 1.12 in its entirety, they should have just released all the content and mechanics at launch as it was the time of 1.12.
Oh I know.
I just think fixing the exploits is a much easier way to go. . Because no matter what being on a locked layer is going to have issues as we both pointed out.
Fixes needed
- Debuf (which they already started to introduce)
- Only able to switch layers in group when in a rest zone.
- For PVP servers, as it does not matter as much to PVE, contested area layer hopping restrictions.
That sort of thing. Let’s hope we see something deployed come next stress test.
1 Like
Do you want layering to last longer? Because that’s how you get layering to last longer.
Where in the hell did you get the idea I want that from? It’s more restricted version of layering that is more dynamic and solves the same issue layering was intended to solve, it will prevent dead servers after the initial rush of tourists leave with eventual merges as pops stabilize. Could it last longer than layering? Sure, but again the tradeoff is zero potential for abuse to avoid or exploit game mechanics.
3 Likes
Going to have to define reasonable if we are to keep this one going.
Spamming anti layering threads isn’t reasonable. Well articulated thread that illustrates awareness of the reason layering is Blizzard’s choice solution is reasonable.
But, knowing that unreasonable people will also reply does not mean Blizzard didn’t take notice.
Blizzard doesn’t make blue post replies all that often, reasonable or not.
Dynamic layering will automatically collapse when possible, because it can expand if it did it too early and haven’t officially “disabled” it.
Static layering will require a human to flip a switch during maintenance (really run a script) and suddenly, the servers are merged.
It will guaranteed last longer than Dynamic layering, because humans will second guess themselves, and once its done, it can’t be undone. Therefore static layering will last longer than Dynamic layering.
The AH is shared. So your only “gain” is no layer hopping. The resource increase is still there, but you’ve restricted anyone from talking to two thirds of the community they’ll be dealing with for the next 2 years.
Fixing the exploits is a far better plan than your arbitrary locking. Cooldowns and rested zones do enough to limit the exploits. If necessary, have a default layer reset after leaving groups too.
Agree to disagree there. What specifically would you suggest to fix the exploits with layering?
Only there won’t be any layering to last longer from a player standpoint. Each “layer” would act as it’s own server, completely isolated from the other layers on the server, with the exception of shared name databases, with no way for players to jump between them. Just give players an option of which layer to join so they’re with their friends(just like choosing the server) and that’s that. Merge if the population drops low enough by phase 2 and beyond, leave them alone if it doesn’t.
2 Likes
Meaning that the “one community” would remain segregated for far longer. That’s the point. The longer they last, the more likely a “realm merger” effect becomes.