diff options
Diffstat (limited to 'pipermail/nel/2001-February/000241.html')
| -rw-r--r-- | pipermail/nel/2001-February/000241.html | 83 |
1 files changed, 83 insertions, 0 deletions
diff --git a/pipermail/nel/2001-February/000241.html b/pipermail/nel/2001-February/000241.html new file mode 100644 index 00000000..3ee5ec68 --- /dev/null +++ b/pipermail/nel/2001-February/000241.html @@ -0,0 +1,83 @@ +<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN"> +<HTML> + <HEAD> + <TITLE> [Nel] Dynamic load balancing?</TITLE> + <LINK REL="Index" HREF="index.html" > + <LINK REL="made" HREF="mailto:archer%40nevrax.com"> + <LINK REL="Previous" HREF="000240.html"> + <LINK REL="Next" HREF="000214.html"> + </HEAD> + <BODY BGCOLOR="#ffffff"> + <H1>[Nel] Dynamic load balancing?</H1> + <B>Vincent Archer</B> + <A HREF="mailto:archer%40nevrax.com" + TITLE="[Nel] Dynamic load balancing?">archer@nevrax.com</A><BR> + <I>Wed, 21 Feb 2001 11:21:23 +0100</I> + <P><UL> + <LI> Previous message: <A HREF="000240.html">[Nel] Dynamic load balancing?</A></li> + <LI> Next message: <A HREF="000214.html">[Nel] RE: Dave Null</A></li> + <LI> <B>Messages sorted by:</B> + <a href="date.html#241">[ date ]</a> + <a href="thread.html#241">[ thread ]</a> + <a href="subject.html#241">[ subject ]</a> + <a href="author.html#241">[ author ]</a> + </LI> + </UL> + <HR> +<!--beginarticle--> +<PRE>According to Thierry Mallard: +><i> On Tue, Feb 20, 2001 at 11:21:38AM -0300, Nahuel Greco wrote: +</I>><i> > [... case of zone server failure ...] +</I>><i> > Then, you has the map divided in zones,... if a geographic server dies, other +</I>><i> > will be taking the zone that belong to the died server? .. +</I>><i> +</I>><i> The problem here will be persistency (is that english ? ;-). All the data +</I>><i> currently in memory on server A, which is failing, will be lost for server B +</I>><i> which is supposed to take over. +</I> +That's what you get with a purely geographic design, like EQ/UO design. +When a zone fails, you get kicked out, everything you gained since your last +save is lost, and when you return, every single critter has respawned. All +state is lost, since all state was located in a single process. + +With a functional approach, all game entities are "holographic" +regarding to storage. They're a kind of "cloud" of various agents located +on various services. If a geographic server dies, you get suddendly +a kind of fog for a while. Another service starts up, gets back information +about what critters are where from the AI services, where all players are +from the various player services, and what items are grown where from the +ecology services. It takes a little time, but you recover more +gracefully. + +The same occurs if an AI server dies. Those critters that were managed +by that service freezes, but then another service takes them over. They +will probably forget a lot of what they were doing, but they don't "depop", +since the geography service knows that there's that kind of beast +that was located there. + +Of course there are services that are more critical than other. If a login +service dies, all players using it are kicked out forcefully. If the +item manager dies, your inventory locks up completely (but you still look +like yourself, can move around, and talk). + +-- +Vincent Archer Email: <A HREF="mailto:archer@nevrax.com">archer@nevrax.com</A> + +Nevrax France. Off on the yellow brick road we go! + +</pre> + +<!--endarticle--> + <HR> + <P><UL> + <!--threads--> + <LI> Previous message: <A HREF="000240.html">[Nel] Dynamic load balancing?</A></li> + <LI> Next message: <A HREF="000214.html">[Nel] RE: Dave Null</A></li> + <LI> <B>Messages sorted by:</B> + <a href="date.html#241">[ date ]</a> + <a href="thread.html#241">[ thread ]</a> + <a href="subject.html#241">[ subject ]</a> + <a href="author.html#241">[ author ]</a> + </LI> + </UL> +</body></html> |