Dynamic Respawning Vs Sharding


#1

Edit: This post is a discussion on how the Dynamic Respawns and Sharding may affect the launch of Classic. I’m hoping that some of the points I bring up here can help people on these forums think about Dynamic Respawns and Sharding in a new perspective. Even though I draw conclusions, I welcome community input.

Let’s first get something out of the way. What is the goal of creating Classic WoW?


  1. Dev Watercooler: World of Warcraft Classic
    https://worldofwarcraft.com/en-us/news/21881587/dev-watercooler-world-of-warcraft-classic
    All the work we’re doing will ultimately allow us to recreate an authentic classic experience

An authentic classic experience. Maybe that was just a slip of the tongue.

  1. Creating World of Warcraft Classic: Restoring History at BlizzCon 2018
    Timestamp 30:00
    “We wanted to come up with a framework that will help guide us. And first and foremost really is authenticity is our goal. We want to create an experience that feels just like 2006 World of Warcraft.”

They’ve said it twice now. Maybe it’s just coincidence.

  1. Kaivax
    Spell Batching in Classic
    As with many other areas of WoW Classic, authenticity is our primary concern.

I’m seeing an authentic pattern here.


So it’s clear the Classic WoW team takes recreating the original game experience seriously. We’ve got multiple examples of that with them compromising on loot sharing, spell batching, spread out content release and I’m sure more in the future. As far as we know they still want to use Sharding during the classic launch. Now it should be quite obvious that Sharding wasn’t a system in Vanilla and yet they are seriously considering using it.

Let’s just say that they use Sharding.
Put down your #NoChanges pitchforks. I’m one of the hardcore supporters of #NoChanges myself.

So we’re getting some form of Sharding for some duration after launch for some level range. Ok, sure. It’s clearly a Quality of Life (QoL) change. So what I want to discuss is why is one QoL change is better or worse than another QoL change.
Specifically:

Sharding Vs. Dynamic Respawns
I’ll state my opinion and lets discuss this in sections.

The Argument

  • I think that using Dynamic Respawns is a significantly better solution to a Classic Launch than Sharding.

Dynamic Respawns is not just a Pserver thing. Blizzard uses it all the time in modern WoW. If we’re gonna use one future blizzard system (sharding), then other blizzard systems (dynamic respawns) should be fair game. Both systems can be tweaked. Sharding can happen harder in the starting area 1-5, and lessen 5-10. Just as dynamic respawns can be faster in the starting area 1-5, and slower in the 5-10.

  1. The Player Experience
  • Expectation

In Vanilla WoW you did not have players phasing in and out of your shard. There was an expectation that if some shaman was ganking people in STV, and you showed up to stop him, you’d see them. Even if the Sharding system is only used in starting zones the point is clear. There is continuity of the world. Sharding breaks this while Dynamic Respawns does not.

  • Continuity

Mobs respawn. Players get this. Mobs may not all respawn at the same rate (rares taking longer etc) but it’s a core system of the game. In fact different respawn rates are a part of the game. Tweaking a system that is already in the game (respawn mechanics) is a lot more consistent with a Vanilla experience than adding a foreign system (sharding).

  • Server Stability

This may be an issue with Blizzard’s hardware. I have heard that Blizzard uses a cloud type infrastructure so that processing power can be dedicated to shards that need it. I remain 100% unconvinced sharding is necessary for stability. I have seen Pservers launch without sharding with minimal lag, clean spell response times and no crashes. This was ND with ~10,000 players on a single server during launch.

  • Total Mob Impact

A shard is going to create a new instance of all the mobs. Essentially this is going to double the amount of available mobs to kill. Cutting the respawn time in half will achieve the same thing. So let’s run hypothetical numbers. There are 6 starting zones with 3500 players. That’s 583 players per zone not accounting for race popularity. So you make 5 shards resulting in ~117 players per starting zone. Think three raids spread out across your favorite starter area. You’re still going to have 5 minute respawn timers on your sharded mobs if they’re normal mobs, which for Vanilla they were.

So dynamic respawns then. Take your 583 players and cut the spawn timers by a fifth. That’s the same number of total mobs in a given time frame. Those starting creatures are going to respawn every 60 seconds. The thing is you can push this further where with sharding you can’t. Cut the respawn timer down to every 30 seconds, or even every 15 seconds. Bare in mind that every single spawn spot will be camped by 5 or more people with 583 players. To get the same amount of mobs using sharding as a 15 second respawn you’d have 20 shards. 20 shards would result in ~29 people in a starting zone.

I don’t know about you but seeing 29 total players in my shard would be pretty disheartening. Though for the person who likes single player games I suppose that would be a perk.

  • Total Mob Experience

Both systems if tweaked equally yield the same total amount of experience within a given time frame. However once the player counts on the shards dip below a certain threshold, mobs wont be camped and killed on spawn. So this generally gives a leveling advantage for dynamic respawning. Players will level faster if they are just grinding. Keep in mind that we’re assuming servers are full to capacity and new players can’t log in. More experience means players distribute into the world faster lessening congestion.

  • Economic Impact

Let’s break this up into mob loot, profession node respawns, and economy as a whole.

More mobs are going to mean more loot, more vendor trash, more greens. Keep in mind both systems provide players with more mobs. Both systems will make the average player have more spending silver. This means players will have an easier time buying their class skills and gearing in the early game. By gearing I mean getting lucky level 5 greens and buying full white chain mail from an armorer vendor (I am so ready for Classic).

More mobs in both systems means professions such as skinning are going to get an advantage. There will be more cloth for tailoring and basically any profession that relies on mob drops will be easier. So mining nodes and herbalism herbs are an interesting point. Normally nodes are not camped, they spawn and remain up for some duration before being gathered. I don’t know the exact respawn time for nodes probably between 5-20 minutes. Even though both systems can have the same amount of nodes per time frame I’d assume that with more players in a given area, dynamic respawn system, any spawned node would be mined faster than people on shards travelling along resource routes. This means basically all professions will be easier than Vanilla to start with using both systems. With dynamic respawns making mining and herbalism easier than sharding.

The overall economic impact shouldn’t be particularly damaging. The amount of wealth generated as a result of both systems in the starting areas should be fairly low. The earlier these systems are abandoned the less economic impact there will be. If the systems span from level 1-10 I’d predict manageable economic impact. That being the amount of money players are making is keeping up with the amount of resources ending up on the auction house to be bought. If sharding or dynamic respawns extend into the 10-20 range I’d start to worry about item farming.

  • Tourism

Lets break this down into what is tourism, how dynamic respawns addresses this, and server health in the long run.

Players are going to join and leave the game. This is obvious. The unique case for this launch is that there is an audience, BFA players, that will be able to try the game for free unlike us specifically subscribing to play Classic. Blizzard seems very concerned about players trying out Classic and leaving. There is speculation that Blizzard will have fewer servers with higher population caps (5-7k). Even though the game world itself was never designed around those higher populations. This is so once players try the game and find it’s not for them then leave you’re left with, hopefully, a healthy normal server population around 2.5-3.5k. I don’t think Blizzard is necessarily wrong to consider sharding. I do think sharding is a solution, but not the ideal one.

Dynamic respawning is scaleable and toggleable just like sharding. It can scale to the number of players, it can be disabled when certain criteria are met. This means dynamic respawning can balance the original Vanilla world for higher population caps to a point. Specifically the starting zones just like how Blizzard wants to use sharding for those zones. However this is where sharding beats dynamic respawns unlike total mob impact. Depending on the server size to balance the two systems you’re going to end up with mobs respawning every second after dying and potentially over a 900-1.3k players in a single starting zone. There comes a point where everyone needs to take a step back and realize that, yes, this is ridiculous. It is really important to understand that this all depends on the server population cap.

When there are too many players for the world. The natural solution is to add another server to balance the overall game population. That’s where long term server health becomes a concern. It is predicted that if normal server populations are used, and there is enough servers so that login queues are reasonable ~20 minutes or less, then that’s too many servers. That’s because a percentage of players on all servers will leave as tourists. Resulting in a population that is lower. It is predicted that server populations will lower enough over several months to be damaging to the server. That’s because the players who still want to play are spread out among too many servers. The only way these two systems can address this is by funneling more players into less servers, resulting in higher population caps. People are so concerned about dead servers they’re not considering maxed out ones.

Higher population caps are a problem when the players don’t leave. The game world was not designed around having 4-8k or more players on a single server without making dynamic respawning adjustments or sharding across the entire 1-60 experience. If Blizzard wants to maintain a higher than Vanilla population cap on launch they will face over crowding issues on servers that continue to draw new players or alts. This means Blizzard is unlikely to raise the player cap significantly over the Vanilla population cap. At most perhaps 500-1000 players with the upper limit reaching the limits of the dynamic respawn system. They may lock those servers from new character creations some time after launch and slowly bring them down to Vanilla levels. If the player cap is not significantly higher than the original Vanilla level than Dynamic Respawning is a viable alternative.

  • Conclusion

My conclusion is using dynamic respawns is more authentic than using sharding. It solves most of the same issues without damaging the continuity of the game world. Dynamic respawning is a better solution than sharding on Vanilla sized server populations. Dynamic respawning has the ability to address tourism, but not as effective as sharding on larger populated servers (megaservers).

I’m interested in hearing community opinions on one system versus the other.


#3

My issues with Dynamic Respawns is when it breaks and mobs spawn so fast that you can’t loot objects for quests or you can’t safely get out of a cave. I really don’t like fighting my way through mobs and have them respawn on me as I eat and drink and prepare for the next.

They most likely will do sharding up through zones that are 6-12 or up to 20 level zones. It should be gone in a day or so. There isn’t supposed to be enough mobs for you and you are supposed to be working with/against other people to get the quest done. It’s a competitive game, no participation awards.

Thy should have put a box price on it and just have a Q for launch w/ a slowly increasing server population.


(Cheater) #4

I feel like Blizzard would code this a lot better than a private server would.


#5

Two non authentic options.

Sharding changes how people are placed in the world.
Dynamic respawns changes how the world itself works.

I still favor having more servers and have them in a preassigned merging pool for when populations decline. It’s not authentic either but it’s closer than either of these two options.

But if it comes down to those two I will grudgingly take sharding for a while over dynamic respawns.


#6

Completely true. Launch and the large surge of players that come with it is a unique scenario. It’s possible that Blizzard would have more adaptive dynamic respawning in place to mitigate this, but that’s just adding more complexity.


#7

They can’t even fix the apparent dc issue with right click reporting and it’s been a year. When people get muted it drops them from the voice chat and it dc’s them. I don’t think they can create a new system of dynamic respawns.


#8

I’ve thought about that but some may argue that you’re just making two really big shards then. Especially if they share names so you can’t have duplicates. It also doesn’t work if one of the pre-assigned servers is very healthy but one starts going low pop. It will also disrupt the economy significantly on the merge.


#9

They will start putting bandaids over bandaids. It actually reminds me of how my work PC had a VGA to DVI adapter and then a DVI to HDMI adapter attacked to it. I have two adapters on it so i can use a monitor that only has a HDMI plugin.


#10

Yes completely true although I was actually thinking more on the lines of six really big shards that are then merged as necessary. And duplicate names get the unfortunate honor of picking a new one.

And yes the economy will be affected aswell completely agree.
I am not saying it’s a good solution. I just haven’t seen a good option suggested anywhere so I am trying to come up with something that is least bad.


#11

So you favor connected realms? Because that’s what you’re suggesting.


#12

So. I would prefer sharding over dynamic respawns because as you level there’s still going to be huge zone overpopulation in the first few zones which they have said to turn off later for a more authentic experience but to mitigate the entrance experience that gave us opening day of vanilla…

While I want it as authentic as possible if by sharding the first few weeks makes everything run in a VERY smooth manner? I’m down.


#13

Don’t exactly know what you mean by “connected realms” What I mean is that servers 1,2,3,4,5,6 are listed into pool A and they will be merged with eachother on a later stage if necessary. In an example scenario: server 1 retains healthy population and won’t merge. Servers 2 and 3 lose half of the populations and merge with eachother and 4,5,6 are basically dead servers with only 1/3 population remaining so they merge with eachother. Now we are left with 3 servers out of the original 6.

If this is “connected realms” then yes I am in favor of it.


#14

Dynamic respawns were actually in vanilla…

Anyone who rolled on the RP-PVP server when they rolled out knows…

It was only in the first couple of days though.


#15

Making the assumption that blizzard is even able to house thousands of players in a single northshire abbey or valley of trials without the servers melting down–

–and assuming dynamic respawns would follow the same timeline (couple weeks) and zone logic (1-10) as what has been implied thus far

–under those assumptions, with a significant enough population, respawns go to nearly infinity, right? instant respawns. well that’s just fine- grinding is now more effective than questing. for experience per minute, for gold. no point in doing quests at all until westfall/barrens.

which is fine i guess – compared to the alternative ‘gongshow’ everyone in together with no sharding and no dynamics where the most effective leveling strategy is exploration.

so you’re messing with gameplay either way. take your pick.


#16

^citation needed


#17

I was there… the opening of the Maelstrom server, they introduced increased spawn rates on monsters temporarily


(Espure) #18

I think I’d prefer dynamic respawns honestly.


#19

Dynamic respawn gets misused, giving certain classes an edge too.


#20

An alternate idea alternate for sharding/dynamic respawns would be to have classic servers have a “Launch Day Cap”. Blizz has the ability to create massive amount of servers and they have shown in the past they can merge servers with no issues. Doing this could have allow Blizz to balance a Horde to Alliance ratio as well. Doing so would provide a classic feel of the game without the overwhelming numbers in a start zone, they could even do delayed server joins. Start out server with say 100 population each and after an hour they allow another 100 to create on a server. Yes this creates more work for Blizz but they have proved so well they are willing to find alternate ways to provide the classic we knew. The numbers I provided are just a guesstimate of what I think other players would be able to tolerate. Hell they could start with 50 on each server and allow 15 player creation wave every 15 minutes, I dunno just throwing out ideas here as I am opposed to sharding as well. I would think this idea would cause at least some frustration when trying to start out playing with friends. The main thing though is would you rather have an idea like this or sharding where you would never see another person that could save your life while questing, or fighting back mobs on short respawn timers.


#21

This is effectively sharding. Having pre-designed servers that will be merged after a few days is a solution but one without much finesse. I would assume this would look like the following:

Darkspear Server 1
Darkspear Server 2
Darkspear Server 3

They would then be merged after say a week. There are several questions that would result from this.

  1. What is the cap on these servers? If it’s the same or higher than a normal server what is that going to do to the population?
  2. Will names be shared across all the servers or not?
  3. How will Blizzard merge these server’s AH and items? I’m under the impression that past server merges have players playing together but with their previous server names attached. Such as Rogar-Darkspear and Dumbly-Frostmane on the same server. Is this just CRZ or an actual server merge?
  4. What is the benefit of this over sharding?

Let’s just run a hypothetical here. Let’s say Blizzard wants roughly 3k players on their server. So the first hour 1k players can play. The next hour 2k total, and the following 3k. This is going to increase the queue times for the server significantly. The solution to that is more servers which is what Blizzard seems to want to avoid by using sharding.

It’s an interesting idea that may work better for much larger scale servers in the range of 5-10k.