At the University of Delaware, a PhD requires you to have submitted "a significant contribution to your field", which usually means a first-author paper in a peer-reviewed journal. Sometimes this doesn't happen, usually because either you have a paper in the works or you got scooped pretty late during your graduate work. These are really two separate issues, getting published and getting graduated.
To gain enough data for a complete thesis, usually several years of work is required. This will depend on your committee (when is enough, enough?), your lab (did you start as a tech your first year?/how much help did you get?), yourself (how hard-working are you?), and your luck (did things work out?).
Getting published, at least once, should be done during your PhD work. Whether or not this is in a big-name journal depends on a number of factors. Right off the bat, if your PI has never published in a big-name journal, you will probably not get a project that will lead to publication in a big name journal. This is why it is so important to pick the right PI. It will also depend on your project. Some people have to work for two years, fine tuning some new method, in order to get out a paper. Some others walk into a lab with a method working, a tech to help them out, and then publish in a big journal within 6 months because everything is working for them. This again will depend on your lab, and is something to consider when choosing a PI.
Now as for generating 2 or 3 papers in big name journals in 6 months. Hah. That would be some seriously prestgous/well-funded PI who set you up with a cush project and you are are a lucky SOB. Sure it happens, but it's very rare. If anyone could do that, everyone would be doing it, because publishing several papers a year in big-name journals will make you an extremely successful scientist.
