Ms Freak
Amarr
NCN Corp
   |
Posted - 2007.01.11 21:35:00 -
[1]
 Originally by: VeniVici
Originally by: Dark Shikari
So clearly you know more about coding and database management than everyone at CCP?
They probably "figured this out" 3 years before you started playing, and have already started to work on it.
Funny how just asking a question of CCP implies that I think I know more than them. 
All I'm asking is if CCP is even looking into other database architectures? Sure at the time the choice may have been clear, but over the last 5 years database technologies have drastically changed, and the choice of architecture made at the start might be far from an effective choice now. The option is there to switch architectures, it does cost time and money, but so do massive hardware upgrades.
CCP knows better the cost of changing architectures, the feasibility of doing so, and the viability of this on performance issues. It's for them to decide but I also know companies tend to take very visible steps to help with performance in order to gain publicity more so than performance. Changing the database architecture is a very non-visible change. It may help performance tremendously but is often not something that can be used to get more publicity. I just hope CCP isn't limiting its solutions in this way.
IMHO You would be correct in assuming thier architecture is the problem, from the data dumps it would imply that there are simply too many tables with too little effiecenty and a sheer record volume issue with the database.
However i would LOVE to know about these so-called "drastic" changes that have happened in database architecture? SQL Server and Oracle have been running on thier repsective engines for more than 5 years now, the only thing that have changed is SQL2005 < SQL2000 and Oracle got some management upgrades (btw: the geo-spatial stuff sux0rs)!?
EvE's issues are more likley due to record volume vs. activity levels rather than anything else. During the early days hardware would have been the problem. once the hardware was sorted the problem transfered to sheer volume forcing over-load on the architecture.
But as many have said before we can only speculate on thier problems. I must agree that every "Upgrade" we get, be it hardware, software, external network stuff or whatever only always seem to result in crap performance for 8 weeks. We finally get some nice performance and then another "upgrade" happens and the cycle repeats.
More "Pre-Emtive" thinking is required rather than "Re-Active".
|