[MUD-Dev] Benchmarking MUDs

Jon Lambert tychomud at ix.netcom.com
Tue Sep 11 02:38:15 New Zealand Standard Time 2001

Adam Martin wrote:
> From: "Ian Collyer" <i.collyer at ntlworld.com>
>> Prime targets for benchmarking MUDs must be...
>>   - Memory
>>   - Processor
>>   - Bandwidth
>>   - Disk space
> That only covers the results from the server view of life. Client
> view of a MUD would be interested in extra things:
>   - Latency (marginal, as opposed to infrastructure-imposed
>   e.g. the inherent latency of my modem dial-up)

I think the most useful and practical measurement is a graphic where
one axis is user transaction response time and the other is number
of users.

But of what sort of transactions?  An average of all types is
useful.  And subsets of all transactions might be created but would
be less than useful depending on whether they are even supported at
all or whether the subset of transactions is even used all that
often on a particular mud.  (i.e. movement commands in a mush)

Which is why I'd still go back to average user command response time
as being the only common measurement.

All other benchmarks are probably only useful in tuning a particular
species of server.  If the idea is to benchmark different species of
servers then there's very little common ground outside of user

I've got a number of hooks in Tychomud that are design to measure
disk usage, memory usage, function timing, and command execution
time.  I test user response time off another box on my local
Ethernet so as to not include internet latency and contention on the
server box in the measurements.  The only trouble is producing
accurate timings with large numbers of users without also including
contention on the client box trying to simulating x numbers of

--* Jon A. Lambert - TychoMUD        Email:jlsysinc at ix.netcom.com *--
--* Mud Server Developer's Page <http://tychomud.home.netcom.com> *--
--* If I had known it was harmless, I would have killed it myself.*--

MUD-Dev mailing list
MUD-Dev at kanga.nu

More information about the MUD-Dev mailing list