[MUD-Dev] Architecture (Cell Rebalancing)

[Zach Collins {Siege}]

On Mon, 30 Jun 2003, Peter "Pietro" Rossmann wrote:

> Is there any better solution to this? As for example - triangulary
> shaped cells would have always only one neighbor on an edge? (if
> moving vertexes as if glued)

Being inexperienced with this sort of thing, I have to ask: how does
the algorithm know the physical layout of the equipment it's using?
In other words, if the cell has a square shape in the data
structures, does that translate directly to a square shape in the
physical network grid?

If so, you're running a very interesting net, and I want to know how
it works.

If not, I would suggest simply measuring average ping times between
units in the cell, and offering nearby Cell Handlers (also defined
by lowest mean ping) some of the overload's higher-ping connections
on the assumption that a high ping generally means longer physical
distance.  However, a connection should probably track its own ping
time, so that a record exists when it's transferred which will note
whether its ping was improved by the swap.  That record can then be
used to help reduce thrashing and shift connections to the best
available cell.

--
Zach Collins (Siege)
_______________________________________________
MUD-Dev mailing list
MUD-Dev at kanga.nu
https://www.kanga.nu/lists/listinfo/mud-dev