Havlentia Castigatrix
The Avalon Foundation
   |
Posted - 2007.09.12 15:38:00 -
[1]
Any TL:DR; previously acknowledged. Don't bother.
 Originally by: Aphroditi
It's a miracle we have EVE running as is.
As someone who works in this kind of stuff every day, with concurrent problems and 'whining' from my customers, I have to tell you that reliance on intervention by metaphysical abstract homocentric concepts is not something that we consider very often.
Incidentally, we don't tend to call the complaints from customers 'whining', because they have this, I dunno, 'valid' concern that they're paying money for something that isn't available. When you contract to large manufacturing, you tend to pay penalty for the amount of time that a line is down for for lost earnings. Likewise making the excuse that this is 'just a game' trivialises the amount of recreational time invested by people.
Originally by: Aphroditi
EVE is HUGE game compared with ANY other game around.
No, it's ambitious. It's also extremely proud of the fact that it's single shard and can handle a large number of concurrent users. Except it can't, not always. The problem is getting worse, not better, and the whole 'need for speed' has been diluted by throwing hardware at a problem that could be in the software/hardware/OS layers. We're being told that they're looking for the problem, but with no real resolution. On top of this, the database strain has been mitigated a number of times by introducing the API and deferred updates to certain tables.
If the software is not up to the strain, then that needs to be fixed. If the hardware isn't up to the strain, then that needs to be fixed. That's what engineers do, we fix things. We don't cross our fingers and pray. Well, mostly not.
Originally by: Aphroditi
Get a life.
I'm an IT consultant. This is my life. :P
Look, I appreciate the effort that CCP put into this, especially from the guys at the lower end of the pay scale. I've been there at 3am watching the bootstrap come up and hoping that it'll work this time around, but in my second decade of doing this stuff I've come to appreciate the value, to my sanity if nothing else, of taking a step back and carefully following a problem through to it's conclusion, and also putting measures in place to mitigate such problems. That's why backups and UPS' exist, even if the entire industry is cutting corners on such things.
Originally by: Aphroditi
A database to support 200,000+ users (30k online)
I have a couple of DBs with over a million users and quite a bit more than 30k online, but nothing like the sheer transactional bandwidth. This is going to be the problem.
-----
This space left intentionally blank |