Patch 7.2... Servers are still popsicle sticks and Elmer's glue

[Daralii]

A: That's called "anecdotal evidence", sir
B: SE's own Lodestone post disagrees with your assertion.

The amount of times this has happened lately is unacceptable. And no, "just get a VPN lol" isn't the answer. This is a severe technical issue that reeks of cost-cutting measures. Wouldn't surprise me if they didn't go with the cheapest possible company to run and maintain the servers.

They went with the cheapest Japanese company they could for the NA servers, which is an obvious recipe for routing disasters.

[Auro_Seldaris]

They went with the cheapest Japanese company they could for the NA servers, which is an obvious recipe for routing disasters.

Color me surprised by that shocking information. Absolutely gobsmacked. /s

Can't remember what person here wrote it, but someone said "Japan used to be 20 years ahead of the world in the 2000s when everyone else was in the 80's, in the 2000's when everyone else was in the late 2000s, and 20 years behind everyone in the 2020s" Very much feels like it, on a tech savviness level now.

[Daralii]

Color me surprised by that shocking information. Absolutely gobsmacked. /s

Can't remember what person here wrote it, but someone said "Japan used to be 20 years ahead of the world in the 2000s when everyone else was in the 80's, in the 2000's when everyone else was in the late 2000s, and 20 years behind everyone in the 2020s" Very much feels like it, on a tech savviness level now.

Just for the sake of also having evidence, here's a thread from 2023 with some records of the awful routing and the packet loss that results from it. The company is Nippon Telegraph and Telephone, which for some reason is federally owned.
https://forum.square-enix.com/ffxiv/threads/482155

[TaleraRistain]

Is this NTT again? Different ISPs seem to have their turn having issues with NTT nodes. A while back it was AT&T so there were a swathe of us being affected. Before that I think it was Verizon. I don't know if SE only has them available to deal with or what but if they do have other options and this is an NTT issue again, I would hope they would consider them.

[Luca_]

"Primal PF is dead and/or trolling, I guess I'll see if Aether is open yet."

"Oh sweet, I can actually transfer there!"

...Immediately before they announced DC travel was broken last night. There's still no way to leave the DC travel queue btw. I just had to call it and go to bed.

The server issues are getting frustrating, to say the least.

[Kaurhz]

Combination of things, but honestly I would say that the biggest issue is Square Enix servers, they do a lot of band aiding, versus direct maintenance or upkeep to actually maintain quality services, and they don't really hide it either, e.g., aggressive login buffers.

Also to my knowledge it is actually very rare for a transit/ISP node to get directly targeted, as these are usually more robust than endpoint nodes, e.g., load balanced, and more redundancy behind it, versus what you would expect to have at an endpoint node, and not to mention transit nodes will typically have a lot higher bandwidth. So it is more practical to attack a specific world/lobby server.

A reason why usually for instance, a DDOS doesn't affect everyone on the server or specific DC, is because to my knowledge the way the infrastructure works, is you have the main lobby server for each datacenter, which will be used for queuing/buffering, fetching main character details (connected world etc.,), and then that would connect you to a specific IP address for that world(s) to maintain your session, and it might be that 1 of those IPs is disrupted/DDOS'd, so it would only affect those connecting to that IP/'server', and then everyone not connected to it would remain fine (even if they are connected to the same world)... A common misconception make is that the entire DC/Lobby only has that single IP/node.

PSN I think do something similar in that they have the authentication/lobby server, and these authenticate, fetch details etc., then it offloads you to a different IP to actually maintain that session... This is why you can sometimes remain connected on PSN despite other folk not being able to connect, or being unable to reconnect if you logout willingly.

[JohnConstantine]

No, it is not. It's not happening for me, therefore it's not the servers.

If it was the servers, it would happen for everyone universally. If it's a node on the way to the servers, it would only happen for some people.

I've also been doing content in the game and the amount of people actually saying they have this right now is barely a thing like it was some months ago, since you can tell by the amount of people getting random disconnects and lag in PUGs, and I've just barely seen such a thing happening recently despite doing Extremes a lot with random players.
It doesn't, because the lodestone post says it's Crystal-only and OP is on Primal. This is a rare case where the infrastructure setup was actually good by isolating the issue to Crystal. Were it more intertwined, the entire game would probably be unavailable.

Does it matter?

The ORIGINAL SIN is that they went with NTT at all. They took the CHEAPEST option to host their servers. They still stuck with NTT when they "upgraded" their infrastructure.

This is like saying: "well, it's not my fault that the building collapsed and killed 10 people. blame the earthquake"

Yeah, well, OK, you chose to build in an earthquake zone in the first place because it was very cheap.

[brinn12]

I once saw around 20 people disconnect in Gridania during a technical issue that was acknowledged by SE, while only me and one other player were unaffected. Claiming the servers are fine just because one person has no issues is technical ignorance. The statistical fact is, NA servers experience more problems than those in Europe or Japan.

[Jeeqbit]

You're never beating the shill allegations bro.

Perhaps you misunderstood. I'm not saying their server infrastructure is good or that their code is any good at making use of it, but that OP's specific issue is not something I'm seeing reflected among players generally when I join parties in PF and DF every single day, which is normally the case when it's a general problem affecting a significant amount of players.

For example, some months ago you'd see average people in PF randomly disconnecting or taking it in turns to, or the entire party saying everything went "slow" then caught up and reporting lag a lot. That's the sort of thing you'd expect to see when playing the game if it was affecting more than just OP, especially in Extremes where the mechanics could go very wrong due to that.

And this thread isn't actually about Crystal it's about OP's specific issues.

The ORIGINAL SIN is that they went with NTT at all. They took the CHEAPEST option to host their servers. They still stuck with NTT when they "upgraded" their infrastructure.

I don't really know the details of what they chose or if they chose NTT and what the alternatives are, but given the apparent "political issues" that I think SE said is affecting NTT, it would certainly be logical to change it if they can.

you have the main lobby server for each datacenter, which will be used for queuing/buffering, fetching main character details (connected world etc.,), and then that would connect you to a specific IP address for that world(s) to maintain your session, and it might be that 1 of those IPs is disrupted/DDOS'd, so it would only affect those connecting to that IP/'server'

If you do a tracert it hops through another IP to get to the lobby server. Similarly, it hops through another IP to get to an individual server on Aether, so attacking the one it hops through probably knocks all 8 servers with one stone. Unsurprisingly, the one it hops through is not responding to ping requests because they are trying to prevent things like that, but I imagine that if the attackers don't know the IP there are only 255 likely candidates and we know it won't be a lot of them due to their use as lobby and server IPs.

[sindriiisgaming]

Ever since this patch, for myself and others in my general region, the lag has been completely insufferable. Common disconnects, >1 second lag, and rubber banding everywhere. Happens in and out of instanced content too.

I've dealt with hardware issues in the past (having to re-instance during Abyssos since my internet would log me out every 2 hours on the dot), and given how widespread these issues are in my circle, as well as my excellent connection on other games, I'm lead to believe that these issues are in fact with the servers. I haven't seen a 2002 since EW dropped, and now it's several times a week? Great!

TL;DR servers suck, everyone knows it. Patch made it significantly worse.

This is your internet provider
i was with BT for a while and had all these same issues, you should maybe look into finding a better company to provide your broadband