Important Information Regarding North American Data Center Relocation

Quote:

---

Originally Posted by LunaFaye

I don't agree with moving the NA data centers to the west coast, something like Texas would have been much better, but when did we have the last major earth quake over here? Right it been some years, it's happening right after the servers are moving is most likely not happening. On the east coast you would have other disasters to deal with, either way I think the best solution would be the middle instead of limiting ourselves to black and white (east and west coast) only.

---

middle US = tornado ally.

heck A decent size tornado hit New orleans and baton rogue a month ago and we still dealing with the flooding that happen mnths prior.
1 tornado even hit behind house when my g.f was home alone. Her first tornado -_-;; lucky my job sent me home.

Quote:

---

Originally Posted by BigRed5392

middle US = tornado ally.

heck A decent size tornado hit New orleans and baton rogue a month ago and we still dealing with the flooding that happen mnths prior.

---

As a life-long Texas native I'd have to note that it's extremely rare for tornadoes to cause major damage to structures that aren't housing, billboards, or stores. The only instance that comes to mind is the Tornado in Fort Worth 17 years ago that blew out a lot of windows downtown. All of those buildings are still there though, they turned the worst damaged one into condos. Texas can have flooding issues too, but that's mainly in rural areas.

A major plus for here versus the East coast is that we don't get hurricanes. They're always tropical depressions by the time they reach the DFW metroplex.

Quote:

---

Originally Posted by BigRed5392

middle US = tornado ally.

heck A decent size tornado hit New orleans and baton rogue a month ago and we still dealing with the flooding that happen mnths prior.
1 tornado even hit behind house when my g.f was home alone. Her first tornado -_-;; lucky my job sent me home.

---

Natural Disasters are pretty low on the totem poll when compared to the higher priorities needed for Data Center planning. No matter where you go, there is a threat. SE also has experience in dealing with them as the Company is in Japan. XI DCs have been their since 2002, besides the XIV DCs. Dealing with several Natural Disasters, the Company mostly dealt with Power Issues with the majority of the events.

Yet you do not see NY, Dallas, St. Louis, and Chicago. Major Backbone hubs with thousands of Datacenters and servers closing their doors. In fact over the years they are constantly improving those services and investing.

Quote:

---

Originally Posted by WarbossGrim

So cut Europe and Oceania out of it, because they could just as easily put them in Europe and Japan datacenters respectively.

---

Except the Aust-Japan link is frequently cut (that same damn subsea cable seems to get chopped fairly regularly) at which point you're routed via the US to get to Japan, making it even slower.

Secondly, the majority of players on the JP servers DON'T speak English and would rather not socialise with them because of the language difficulty.

Thirdly, those servers where Australians and English speakers are concentrated are already full, so moving there is not easy.

Still, as the US servers are already about as far from us as can be, anywhere will be an improvement.

Quote:

---

Originally Posted by Aria_Sevenpeaks

Except the Aust-Japan link is frequently cut (that same damn subsea cable seems to get chopped fairly regularly) at which point you're routed via the US to get to Japan, making it even slower.

---

Our connections to JP are routed through the infrastructure (some of which is not very robust) of four different nations in before they hit Japan, involving multiple sub-sea cables. It's not the same cables going down all the time it's different ones. The result of which is a highly variable connection. Sure on good days Australians can get maybe a 150ms stable connection but most days are not good days.

I have characters on both Tonberry (JP) and on Faerie (NA) and I can guarantee any NA based player that my connection to the NA data centre far more stable. A fast connection to games is not the only thing that counts, a stable connection is extremely important. My connection to NA is 200-250ms most days. My connection to JP bounces between 150 and 900ms most days. It's easy to adjust for latency when the latency is going to be predictable. When it's all over the place it's hell.

As Aria also pointed out the vast majority of JP players don't like to play in English - their English I dare say is about as good as my Japanese limited to basic words and I can only read hiragana and katakana, the only kanji I know are the symbols used for dates. The queue for daily roulette on Tonberry is 30mins + and that's as a healer. If you queue for specific dungeons you have to be prepared to wait for hours.

If SE believes you ideally need a connection of not much more than 200ms to play this game then they absolutely should take Oceania into consideration when placing the NA data centre. Japan may be physically closer to Aust/NZ players but the infrastructure is tenuous making NA a viable choice.

Many MMOs have moved from east or west coast servers to places like Austen and Dallas in the last 5-10 years and in my experience SE moving the servers here would make the game completely unplayable. As I said previously I currently get 200-250 to the NA data centre, in Canada, on the east coast. 200-250 is pretty normal for me connecting to game servers on the east coast. West coast it's around 150ms. Games that have moved to Dallas and Austen in recent years it's 400ms. Why so high, well apparently the infrastructure in the US from coast to internal locations like Texas is pretty terrible. While this can be reduced to about 300ms with ssh tunnelling or a proxy service like battleping and wtfast, this is still higher than what I get to the servers currently without such a service. Nevermind the fact that using such a service can result in a false positive for bans and a wasted few weeks or months trying to get your account back.

People saying NA servers should only be for NA players forget the fact that Aust/NZ players are majority English speakers who are unlikely to have Japanese as a 2nd fluent language (for the record my 2nd language is French and I also speak Spanish so NA servers are a great place for me) and fail to realise that our connection to Japan is not a singular cable straight from here to there - on the other hand our connection to the US is functionally a single cable.

Quote:

---

Originally Posted by Lozza

snip.

---

This is not my experience at all. I played on Sarganatas from 2.0 to 3.1, then transfered to Tonberry until present and have a far more stable and consistent connection. I've had very little trouble here playing from both TAS and Sydney.

I'm guessing your queue time is queueing while only selecting 'English' - which is the first question most mentors ask when people ask about queue time in novice chat. I gt 5-15m queues as a MCH. Off-hours are obviously different, but that's the same on all servers. The 'vast majority' of both English and Japanese players are similar to all languages: Hello at the start. Goodbye at the end. Language plays no role here.

Quote:

---

Originally Posted by Pence

This is not my experience at all. I played on Sarganatas from 2.0 to 3.1, then transfered to Tonberry until present and have a far more stable and consistent connection. I've had very little trouble here playing from both TAS and Sydney.

I'm guessing your queue time is queueing while only selecting 'English' - which is the first question most mentors ask when people ask about queue time in novice chat.

---

For basic dungeons auto translate works fine, however I would never queue for a dungeon I haven't done before in Japanese, the language barrier is simply too great. As for your connection being more stable I am curious if you are running battleping or wtfast (or similar service) these services can often reduce the variability in connection. I haven't run with them for a few years now, but then my connection to the NA servers is surprisingly good and I'm not willing to risk another ban.

Quote:

---

Originally Posted by KokonoeAiyoko

Any chance this means the end of Level 3 issues?

---

This, is the location going to end up using level 3 routing? that has to be the most important thing. If you say the game is supposed to be playable on 200 ms ping, why is there significant lag issues with a current 40 ping to Canada?

Why are europeans complaining? Don't they have their own servers now?

Quote:

---

Originally Posted by Ama_Hamada

This, is the location going to end up using level 3 routing? that has to be the most important thing. If you say the game is supposed to be playable on 200 ms ping, why is there significant lag issues with a current 40 ping to Canada?

---

Doesn't matter. The current issues are more involved than one backbone provider who has been blamed countless times for issues outside their own control and not even within their own infrastructure by the playerbase.

SE has been working on this DataCenter behind the scenes for about a year based on recent interviews. SE is now admitting the current NA DC cannot expand due to physical & power limitations as well as having overvoltaging issues which directly affects the players.

- The current NA DC cannot expand to reduce load limits on the current hardware that is already taxed. This directly effects connection stability.
- The NA DC is the weak point for further game expansions & system improvements to include SB, which affects all DCs. (Why there is a relocation & new DC targeted for May release)
- The Dev Team has been restricted to implement possible improvements and/or enhance the user experience that require server loads beyond the current already taxed NA DC limitations, however the other DCs can handle the additional loads and further expand if required within their current locations.

When a DataCenter's servers & network hardware is overtaxed on loads & on power, the end user experiences "lag". A similar example for the end user would be running their own CPU/GPU at 100% load and experiencing the slowdown caused as the processors are bottlenecked, unable to process instructions at its normal speeds.