
Resource usage (network bandwith, cpu load etc) by MUD playing
--------------------------------------------------------------

	This file contains a - partailly re-written - response that
	I sent to a person who asked me about this issue in 1991.

There has been quite a lot of discussion going on in different places
- various news groups on the net, among others - about MUDs, resources
and usage.
Most of it hasn't been very informative - even if both the flame wars
and the philosophical debates about rights and property can be rather
entertaining when watched from the outside. I don't think there has been
any generally agreed-upon conclusions, and unfortunately I don't have any
pointers to any papers or so, except maybe some general texts about
networks and computer communication.

I'll try to give my personal thoughts about this issue, though.

In some places MUD playing has become a problem, but as far as I have
understood it, the problem has usually been terminals - i. e. MUD players
using the terminals or workstations or whatever to play MUD when others
could have used them for real work - and not the load on the different
networks.

I'm not absolutely sure of what I'm saying here - better consult a net
guru about it - but I don't think a MUD player takes up much of the
network capacity - neither on the local net or on the "bigger" net
connecting him to the MUD game. The amount of data being sent between
the client and the server is not very great really - the client sends
a line whenever the user types a command, and the server sends the output
text from the MUD game, and then we must add the data used by the net
to wrap these in - maybe 50 bytes or so for each package on a wide network.

And each package contains a whole command line (at least when running
in line mode, as PMF does), and in the other direction at least one
line of output text.

Compare this to the traffic on an Ethernet with diskless workstations,
or with the ftp traffic going on on the Internet, and you'll find
(at least I think you'll find) that the MUD traffic is negligible.

But the Internet is a large and complex aggregation of network hardware,
connected together by various kinds of gateways, so strange things could
happen. In special cases the capacity of the network, or of a certain link
in the network, could of course be a limiting factor - like the
(mythological?) 9600 baud link connecting the continent of Australia
to the rest of the world.

At the other end of the line, on the system running the MUD server,
the situation can be worse. The strain - both in terms of cpu load
and disk usage - can be a real problem, at least under some circumstances.

The problem of terminals being occupied by MUD players should (in an ideal
world) be easily solved by some minimal common sense among the players
themselves, maybe with the help of an explicit "work goes before games"
policy.

I don't know the nature of your computing facilities (one big mainframe
with terminals? diskless workstations? PC:s?) but I don't think a few
people running a client program like PMF could cause a significant
or even noticeable increase in the load of your systems, just by the
processes, and as I said above, network capacity shouldn't be a
limiting factor either.

For the undergrad education here at the University of Linkoping (in
Sweden) we have around 100 diskless workstations (Sun SPARCstation 1)
and 4 or 5 file servers. Some of the students use these workstations
to run PMF and play a MUD game run by the local computer club (on
one of their machines, not the university's), and I haven't noticed
any problems at all.

                Padrone, April 3, 1992
