Cart
[Home] [By Thread] [By Date] [Recent Entries]
"Al B. Snell" wrote: > > On Fri, 13 Apr 2001, Stephen D. Williams wrote: > > > No, the difference is that after connection, the rest of the > > session/conversation can be handled by another thread/process/cpu with > > TCP. You can't do that with UDP since there is no way to know at the > > kernel/driver level what logical process an incoming UDP packet is bound > > for. > > Oh, you mean once conversations have started and you want to maintain > state? Well, either keep the table of > which-process-handled-which-conversation in the master process, or use the > master process purely as a broker which replies (on conversation > setup) with a destination port to use that points to the correct child. In > practice, though, if you actually want a stateful conversation like that, > straight UDP probably *isn't* for you! You Can dynamically assign a UDP port that the rest of the conversation can proceed on in a separate thread, but that limits you to 64K concurrent conversations. TCP has no such limit; it can have 64K conversations with Each remote IP address. Presense/IM (i.e. Buddylist/AIM/ICQ/Jabber) is a good model where it has been generally agreed that a long-lived TCP connection is needed. The communication model is the first widely used (as in number of actual end-user/computers) instance of async message oriented routing. In my mind, as I've espoused for a long time, it's a great potential platform for other applications shipping generic XML and more. It is in fact a good example of the savings that message pipelining can buy you. > I'd like RDP resurrected. It strikes me that RDP can be implemented in > userland on top of UDP (which is, at heart, a controlled gateway to IP > itself)... that might be worth examining. > > ABS > > -- > Alaric B. Snell > http://www.alaric-snell.com/ http://RFC.net/ http://www.warhead.org.uk/ > Any sufficiently advanced technology can be emulated in software sdw -- sdw@l... http://sdw.st Stephen D. Williams 43392 Wayside Cir,Ashburn,VA 20147-4622 703-724-0118W 703-995-0407Fax Dec2000
|
