Quote:Original post by Oogst
Through my old university I have access from home over the internet to almost all computer science papers, but I would much rather read a book. If I look for articles, I will be sure to miss a lot of subjects that a book might force on me. Also, I like to just read it in the train from front to back, without bothering to look for stuff. Are there any less advanced books that give a good overview of how to do multiplayer game programming?

You'll be out of luck. There's only a handful of books I've found to be rather dated, and the "forcing" aspect was all about some trivial optimizations, such as a full chapter on how to write a really really good allocator for something.

Clustering is not really a trivial subject, since it isn't really an implementation problem, but a design issue of how to distribute your resources over physically separate hardware and solving issues that arise.

Some MMO publishers have published or at least documented their solutions. Going through existing libraries may be helpful as well. ZeroC's Ice framework discusses (and proposes solutions) to almost all problems you'll encounter, from light-weight serialization to versioning, resource replication, administration persistence. It does not cover timing or reliability issues.

Spread toolkit is an accessible messaging framework which may offer some insight into a practical design of message-based architectures.

Forum FAQ lists several game engines, NEL uses standard service oriented approach and covers the RPC, as well as full-async AMI design.

And ultimately, there's a lot of application servers, mostly from enterprise world, which vary in approaches to scalability.

Then there's various distributed computing frameworks, but those tend to make assumptions which aren't viable. Either they aren't intended for real-time use, they assume reliable systems, or they are simply too focused on the nature of problems.

Quote:Unfortunately, this means looking for likely papers on Citeseer and similar places

I'll second that. There's plenty of research, and also real-world examples on how things were done, examining widely different approaches to solving same problems.

Quote:without bothering to look for stuff

That may turn out to be the biggest problem. The domain is somewhat limited, and to a degree closed. So the need for such global and front-to-back documents doesn't necessarily make it feasible to centralize the knowledge, simply because it covers far too many domains.

If you're not already intimately familiar with the underlying protocols, "TCP/IP Illustrated" is essential reading. This may be worthwhile for the high-level concepts, but I haven't read it.

As others have suggested, you're not going to find good books specifically about what you mentioned in the original post. Perhaps mainly because there just isn't a large audience for it.

I was asking about similar things and was given links to these two interesting articles:

http://www.gamasutra.com/view/feature/3094/1500_archers_on_a_288_network_.php

http://www.gamasutra.com/view/feature/3374/the_internet_sucks_or_what_i_.php?page=1

While not books they contained couple of new tricks I had not known before.

I am also interested in multiplayer networking but from the 3d and protocol perspectives.

regards,
Tommi

Try Zyda's site. There is a course with free power point. Get his book if u want a hard copy.

http://gamepipe.usc.edu/~zyda/index.html