[Home] [Groups] - Message: [Prev in Group] [Next in Group]

nu.kanga.list.mud-dev

25726: RE: [MUD-Dev] Is database access a bottleneck? Prevayler.

[Full Header] [Plain Text]
From: "Damion Schubert" <damion@ninjaneering.com>
Newsgroups: nu.kanga.list.mud-dev
Date: Thu, 12 Dec 2002 00:52:56 -0600
References: [1]
Organization: Kanga.Nu
From: ra-whiteside@attbi.com

> Now, being a gaming enthusiast (though not a game designer or
> developer), it occured to me to wonder if this technology would be
> useful in MMORPG games, assuming it indeed works as advertised .

I'm the designer, not the database guy, but in most implementations
I've seen, the database is not read at runtime.  It is loaded upon
bootup (and/or first access), and then written to at runtime for
backups, but most of the necessary game data is kept around in
memory.
 
> So, my first question is, "What is the bottleneck in scaling
> MMORPG games?"  Why do we have worlds with 2-3K players instead of
> 200-300K players?

As Matt said, content is the biggest reason.  There are other
reasons (cross-server traffic, the threat of congregation, etc) but
as a designer, the real scary part would be making a world 100 times
the size of Everquest without ending up with incredibly pointless
and stupid randomly generated dungeons.

Another reason is namespace for unique names.  Yet another reason is
that multiple servers means that more groups can be the winner of
your biggest metagame.

In licensed properties, such as an Ultima or Star Wars, you could
HAVE the content, but people would still want to go to places they
know.  Meaning, you could have a million randomly generated worlds,
but everyone will want to be on the Death Star and the Cantina in
Mos Eisley.

Sorry, it appears I got off on my trademarked '1 million person
shards are bad' rant without meaning to again.

--d


_______________________________________________
MUD-Dev mailing list
MUD-Dev@kanga.nu
https://www.kanga.nu/lists/listinfo/mud-dev