Yes, I know. Which is why I wrote this in a follow-up reply earlier;
Contrary to your contrarian view, if you change the hardware in a LIVE environment, you are risking breaking something in the software, which in turn threatens the service. The whole OS and application Stack on the server is configured for the specific platform. You can't just change hardware in the server on the fly. If something fails, it will be replaced by an identical unit. In a production setting where 24/7 operation is expected, you don't mess with the configuration.we are approaching 3 years, and the 1st anniversary of Heavensward. It just seems to me that now would be a good time for them to begin planning for a server refresh along side the 4.0 expansion; which would make it possible that we would see some of the changes people want to see with that expansion. Given how long it takes for major changes to materialize, I figured now would be a good time to start the conversation.
Changes to server hardware are carefully planned in advance, maintenance is about restoring the system to operation, not stealth upgrades. OS patching is a more continual (and critical) process and with the risks of unpatched exploits on servers, it's routine now to upgrade the OS and patch applications as needed. But the underlying hardware won't change without a good deal of planning, which is why you start planning long before the server hits EoL, or goes out of maintenance.
I'd consider the Eurpean data center an upgrade, since it was in fact an upgrade. Montreal was where the NA/EU data center(s) were based from the start of ARR, although as someone pointed out earlier, the original datacenter infrastructure was not designed to handle as many players as the game got, and had to be scaled up, hastily.
Yes, Indeed. However, as Yoshi-P and others have told us time and again the memory budget for character data is very tight. That is partially due to constraints on the client end, but mainly due to the constant write back of character data to the database every 15 (It may actually be even more frequent than that, I can't remember, but 15 seconds has stuck in my head) seconds to provide solid data integrity and rollback support. You could quadruple the number of options for character appearance, and double inventory, and most clients (including the PS3) wouldn't care. However increasing the number of options for character appearance and doubling the inventory would have a large impact on the amount of data each character comprises of, and is being continually re-written to the database. The current architecture supports a certain amount of data in the character, you have to upgrade the actual hardware to significantly alter than bandwidth to support more data in the character.
Increasing customization options, dye options, glamour options etc... all increase the amount of data stored with a character as well. Sure, weaker clients might not be able to keep up with rendering everything, in which case some enhanced options might not render fully on those platforms, rather like what would happen with a high definition texture pack. But, anything that increased the size of the data set that makes up a character has to be included in the budget for the size of character data that can be handled by the system.
Of course there is a large software component to making this happen, but the hardware is a limiting factor and upgrading hardware raises the bar on what can be done, facilitating the change in software. The two go hand in hand, but it's a chicken and egg. You can't expand inventory without first increasing the capability (bandwidth) of the game servers to rewrite character data continuously. The only alternative to changing the hardware is to completely rewrite the code and database that handles character data so that it is more efficient and/or stores things differently. To me, it would see that the best solution would be to do both, but that's just me.