I tested a server in blitzplus with 150 clients sending/reciving packets to the server about 1 hour without any problem and a really fast response from server...
The math disagrees with you. Bandwidth requirements grow exponentially, as the number of clients grow linearly. Even if you're only sending 64 bytes (the lowest possible size of an Ethernet frame) you'e pumping out 1.36 Megabytes each *syncronous* update. Say you have 5 of those each second (which is low, roughly about every 200ms) you'll be punching out nearly 7MB every second, sent synronously constantly being interupted by more than 1500 context switches?
I don't see how that's even remotely plausible, even if you where doing nothing else, but sending random data. I guess we have a difference of opinion on what constitutes a "fast response from [the] server".
Since when programmers use threads?
Since about 1971.
Couldnt find that info on the Net, maybe its old tech and i mix it with hyper threading..
Yes. HyperThreading is basicly a "poor mans dual core" processor. In order to increase performance in Intels single-core CPU's Intel engineers decided to give each core, two instruction pipelines. This means that while the core is executing instructions in one pipeline, it can load the other with instruction, thus significantly lowering the performance impact of a context switch.
Using one of the .net languages with a SQL Server based product is great for the backend. Either MSDE(based on SQL 2000) or SQL Server 2005 Xpress
No it's not. MSDE support expires ultimo 2007, and doesn't work at all on Windows Vista, and SQL Server 2005 Express doesn't have a scheduler agent. Add to that the 4GB data limit which pretty much eliminates it as a viable option for anything wanting to be "massive" or "persistent". A much better solution would be to get a proper database, like Postgres or Firebird.