Was it a mistake to deprioritize data center travel?

Quote:

---

Originally Posted by -BlueGreen-

True enough, though I'd like to think that securing bandwidth for this purpose is likely far easier than securing servers for additional worlds. Of course, thinking on it, the real challenge would likely be in whatever configuration changes they'd need to make and test out for it would likely require engineers to be on site to run cables and such possibly, and COVID travel restrictions being what they are might actually be the biggest blocker.

---

I think it's more than bandwidth, its latency, you wouldn't want instances in Japan waiting for 500ms+ (and P/L) to write/read data back to an NA DC database. Instances for PvP/Raids are simple since all the data is contained within that one physical server.

There shouldn't be any physical changes required for regional DCs since they all reside within the same physical location, that's more software based as you proposed. I think just based off the restrictions on current home world transfer service, sharing the same data source for multiple regional DCs would still require significant effort.

There's also the limitation of the queueing server that distributes clients, I think they said the current limit is 21k, imagine that spread over 30 servers instead of 10.