[MUD-Dev] Middleware (was re: Sun's Sim Server and Gordon's 10 Reasons (thefirstone :))

[Amanda Walker]

On Apr 5, 2004, at 1:05 PM, Tess Snider wrote:

> I have joked with friends before that there are currently more
> MMOG server middleware providers out there than there are
> companies that could realistically benefit from them.

This is not just true of "MMOG server middleware" :-).

Middleware is an extremely seductive concept, both for providers and
customers.  However, it only really works in domains that are mature
enough for interfaces to have stabilized.  DirectX and OpenGL are
good examples of successful middleware: to my knowledge, no one
bothers to write hardware-specific 3D code any more (driver- or
manufacturer- specific, sure, but I don't see game developers out
there banging bits directly on the GPU the way they used to).  IP is
another successful example: remember all those "tunnel game traffic
over IP" utilities from years past?

However, these examples work because using mature middleware really
is lower cost than rolling your own, no matter how skilled you are.
Game logic is nowhere near that mature yet.

> Besides, anyone who claims that networking, distributed
> processing, server architecture, security, and data storage are
> solved problems is going to cause a lot of folks on this list to
> snicker into their beers.

Yes and no.  On a large scale, many aspects of them are often solved
better outside the game industry than they are inside it, and this
is not simply because games are essentially different from the rest
of the world--there is a distinct "NIH" attitude in the gaming
industry.

On the other hand, each industry does have its own requirements and
concepts.  These are often not obvious from outside, and middlware
(or any -ware) providers often miss this when try to enter new
industries.  MMO systems pose database problems, network problems,
distributed computation problems, etc., but the requirements and
mixes of these really are unique--as are the particular mixes
presented by other industries (library systems, bioinformatics,
ISPs, fleet management, whatever).

> I think that where I see MMOG server middleware coming into play
> is if, for example, you just landed a really hot license, and
> needed to get something to market really fast.  (Boy does that
> sound like an ill-fated venture.)

I think that's a bad advertisement for middleware.

Where I think middleware will come into play is in the areas that
innovative game value leaves behind.  It's already happened with
graphics, and mostly happened with networking, at least at the
transport level (show of hands: who's working with raw Ethernet
frames these days?).

I think that a common object and messaging model has huge potential,
but not yet.  Even the DoD, who can define a standard and just order
people to use it, hasn't managed to make this work yet (though
anyone seriously interested in MMO middleware would probably profit
by boning up on HLA and CTIA and the lessons learned in doing so).

Once the distributed object layer stops being interesting, though, I
think we'll see some interesting middleware.  Intercommunication
between *different* MMOGs is a fascinating idea, and fits in with
the "use the web as a user-generated content model" idea some of us
were tossing around here a year or two ago.  However, this only
works with a uniform treatment of objects, messages, attributes, and
so on.

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