Datacenters: A Plea For Community Building, Not Fracturing

[LaylaTsarra]

Except this won't prevent it from happening again because most players refuse to make smart choices. They'll once again zerg the game at release, causing bottlenecks and crashes.

We'll have to wait and see if it makes a difference. I don't have a crystal ball but what was happening wasn't a good thing.

[Bourne_Endeavor]

Except this won't prevent it from happening again because most players refuse to make smart choices. They'll once again zerg the game at release, causing bottlenecks and crashes.

I'll reiterate one more time. This wasn't a player issue but a SE one. They put an instance ten quests into the MSQ, thereby ensuring virtually everyone would all queue at roughly the same time. Move that instance further down and this doesn't occur. Why? People will naturally veer off. Some are going to blitz the story and skip everything, others will take their time. Some will go level other jobs or do side quests along the way. Speaking of leveling jobs. Guess what also had instances in their first quests: Red Mage and Samurai! It's almost like having three instance extremely early on during an expansion launch is a very bad idea.

For only two weeks? Really? That's acceptable to you?

No it's not acceptable in my opinion and that's not to mention the fact that many after waiting forever in queues of over 3000 to get in only to get kicked and to have to get back into line again. I applaud SE for taking measures to try and prevent that from happening again. One can hope that this will in fact work and I'll guess we'll find out when expansion hits.

If the choice were between what ultimately is a mild inconvenience for 4% out of a given year's total weeks, yes. I would rather have the larger pool we currently have for PF and DF since I intend to play far larger than those initial two weeks. So less people means potentially slower results long term. And those massive queues weren't an issue on other servers. Said queues also died down completely when they did force log outs. So the tools are already there to mitigate the problem. Implement the forced log out immediately and don't put three instances super early into quest chains and maybe the problem doesn't occur.

[LaylaTsarra]

If the choice were between what ultimately is a mild inconvenience for 4% out of a given year's total week

It's not a mild inconvenience it rendered the game unplayable for many people during the prime time of the expansions release. Sorry but your willingness to accept that isn't what the development team is willing to accept and they are making choices to try and prevent it from occurring again.

[Bourne_Endeavor]

It's not a mild inconvenience it rendered the game unplayable for many people during the prime time of the expansions release. Sorry but your willingness to accept that isn't what the development team is willing to accept and they are making choices to try and prevent it from occurring again.

It is when you compare it to the alternative, especially as two weeks out of fifty two isn't exactly a lot. Nor was the game actually down for two weeks. Several hours were lost, yes—more, admittedly, if you were on Balmung and Gilgamesh. But like I said, the issues primarily occurred because of the instance quests they put so early into the MSQ and both new jobs. Many people now are inconvenience far greater as they have give up their static, house, FC or some of their friends. Ironically, if queues do slow down due to a less populated DC, then you'll lose similar hours stretched over the course of Shadowbringers. Making it a moot point regardless.

For reference sake, Heavensward had no such issues despite Balmung and Gilgamesh's population being roughly equal to what they were at Stormblood's launch, if we go by Lucky Bancho's census. What differed? There wasn't an instance where everyone dog-piled onto within the first hour. Furthermore, like I mentioned earlier, these log in issues happened again during Deltascape. And what did Deltascape have very early into its quest chain? An instance. Yoshida even acknowledged he shouldn't have allowed the dev team to go ahead with it.

[Hamada]

It is when you compare it to the alternative, especially as two weeks out of fifty two isn't exactly a lot. Nor was the game actually down for two weeks. Several hours were lost, yes—more, admittedly, if you were on Balmung and Gilgamesh. But like I said, the issues primarily occurred because of the instance quests they put so early into the MSQ and both new jobs. Many people now are inconvenience far greater as they have give up their static, house, FC or some of their friends. Ironically, if queues do slow down due to a less populated DC, then you'll lose similar hours stretched over the course of Shadowbringers. Making it a moot point regardless.

For reference sake, Heavensward had no such issues despite Balmung and Gilgamesh's population being roughly equal to what they were at Stormblood's launch, if we go by Lucky Bancho's census. What differed? There wasn't an instance where everyone dog-piled onto within the first hour. Furthermore, like I mentioned earlier, these log in issues happened again during Deltascape. And what did Deltascape have very early into its quest chain? An instance. Yoshida even acknowledged he shouldn't have allowed the dev team to go ahead with it.

Yeah this is why this DC split is a bad idea, outside 2 weeks, what happens during the other 50 weeks? dead PF? slow queues? PvP not existing on crystal? They did not even explain why this is being done, all we can assume it is based off SE is unable to figure out not to put instances so early in a new expansion where we did not have this issue in HW.

Keep in mind for those that want to blame balmung and the playerbase, SE did completely unlock balmung in SB not to long ago ( I do not remember exactly) but like 4.3? or so? that is fairly recent, so why the sudden turn face making the playerbase it was fine then, to now OMG SPLIT DCS for short cutting to avoid more costly and permanent solutions? Stop bandiading problems!!!!

For people supporting the split for DC, you do realize the amount of active players at any given one time in this game is not very large when you compere it to wow or something and didn't others say they actually fused data centers because they upgraded their tech? SE can't do this why? I do not see an excuse other then "we do not want to spend money but we can milk the playerbase out of money from retainers, app and 18 dollar outfits for one charter." (btw , they had tech excuses for not sending account wide stuff, then proceeded to send account wide stuff)

[aesteval]

This wouldn't have happened if the balmung population wasn't so stubborn about staying on their server despite all of SE's attempts to balance out the population.

Part of the additional datacenter is due to events like Raubahn Ex. Primal was also hit by Raubahn Ex. Raubahn Ex happened without Balmung's influence. The new datacenter is not *just* about Balmung.

[ColorOfSakura]

Part of the additional datacenter is due to events like Raubahn Ex. Primal was also hit by Raubahn Ex. Raubahn Ex happened without Balmung's influence. The new datacenter is not *just* about Balmung.

Exactly. Balmung isn't the problem here - the problem is the game being too popular and needing more infrastructure to support it. Yes, Balmung was moved away from Gilgamesh because they're the two most populated servers on the game in terms of active players and splitting them up is a smart idea in general. But this was something that was a long time coming - Balmung isn't the sole factor of it.

[WaterShield]

Exactly. Balmung isn't the problem here - the problem is the game being too popular and needing more infrastructure to support it. Yes, Balmung was moved away from Gilgamesh because they're the two most populated servers on the game in terms of active players and splitting them up is a smart idea in general. But this was something that was a long time coming - Balmung isn't the sole factor of it.

Except this doesn't actually add any infrastructure. This is a move to more evenly balance people (see load balancing) across the same existing hardware (no new servers being added). That is not much of an investment at all.

Also how do you explain Arena Net's ability to do Mega server tech? Bethesda as well with ESO. Are they just that much better funded than SE? I find that hard to believe.

As many people before me have pointed out, the dated code SE is using is impeding their ability to grow the game. Rather than fix it properly, we are being given a band aid that can be completely invalidated by people transferring back to Aether and Primal.
Shortsighted at best. Purposely negligent at worst. Is it any wonder people think they are using the majority of the money from this game to bankroll their other operations?

[Valkyrie_Lenneth]

Except this doesn't actually add any infrastructure. This is a move to more evenly balance people (see load balancing) across the same existing hardware (no new servers being added). That is not much of an investment at all.

Also how do you explain Arena Net's ability to do Mega server tech? Bethesda as well with ESO. Are they just that much better funded than SE? I find that hard to believe.

As many people before me have pointed out, the dated code SE is using is impeding their ability to grow the game. Rather than fix it properly, we are being given a band aid that can be completely invalidated by people transferring back to Aether and Primal.
Shortsighted at best. Purposely negligent at worst. Is it any wonder people think they are using the majority of the money from this game to bankroll their other operations?

Not exactly, it will add new infrastructure in the form of 50% more lobby and instance servers to accommodate a whole new datacenter (as each datacenter requires its own). So thats at least an improvement there.


As for the "SE is being cheap" part. I agree, but not financially. I believe they are being cheap with time, as any kind of effort to do this would require a massive amount of manhours to rebuild. It would also take the people away from the improvements we ARE getting in terms of QoL for backend and the like. The difference between now and what the game was like at launch is vast, however much people like to say "they havent made any improvements at all!"

[reiichi]

Also how do you explain Arena Net's ability to do Mega server tech? Bethesda as well with ESO. Are they just that much better funded than SE? I find that hard to believe.

Do we know if they make use of something like AWS or GCP? AKA, cloud computing platforms where you pay for the number of instances you need on a dynamic rate and scale (literally clone and boot up) additional servers as needed. It works great for -certain- use cases designed specifically for that.

FFXIV's not doing the same. Instead, it looks like they're using dedicated machines with hard-set numbers of servers and instances for things. There are pros and cons to both. (Usually cost IMO, but one server with a lot of resources can also work out better than a bunch of tiny servers with few resources depending on how the product is designed.) This is why things like housing wards are hard-set and don't just expand based on demand/availability. (Aside from UI constraints like supporting people who play this game in 720p).

Chances are, the new datacenter IS going to involve adding a new rack (or however many they need) of new hardware instead of just spinning up a new lobby server instance and setting which worlds it can talk to. Anything that goes through Content Finder (Duty Finder, PoTD, HoH, instanced solo missions, etc) is shared among all users at the datacenter level. My assumption is that they're not going to de-allocate those servers from Aether and Primal in order to make Crystal's set. They'll get that many more servers/vms/logical partitions in addition. And that means an actual improvement for everyone.