I'd argue that you shouldn't worry too much about preventing plagiarism of your papers. At least in the areas I'm familiar with (math and theoretical CS), it's extremely unlikely to do you much harm.
To be clear, I'm talking about wholesale plagiarism of written text. There can be much trickier situations - for example, if your rival learns an idea from you in a private conversation and then claims you learned it from him - but that's a subject for another question.
Plagiarism happens all the time, on a massive scale, but you don't see it very often because it takes place at the margins: usually in junk journals or conferences, and occasionally in solid but low-prestige venues. This is by design, since plagiarists know that if they attract too much attention, they'll get caught. Nobody submits a plagiarized paper hoping to get lots of citations. Instead, they just want credit for having published something, usually because of some external pressure.
One fear people sometimes have is of being caught in an ambiguous situation, where it's not completely clear who the plagiarist is. In the areas I'm familiar with, this never happens. The community is always quite sure of who is to blame, and as far as I can tell they are always right. This is a valid worry in principle, but it's not worth losing sleep over.
Another fear is that the plagiarist will attract attention and citations that should have gone to your paper. As I mentioned above, that's not likely. These papers are typically almost unnoticed, and if they do get noticed, then that's just a prelude to getting caught.
Plagiarism is still a serious problem, and plagiarists are cheating the system, but in the fields I'm talking about, they are generally not specifically hurting the authors they are copying from.
Putting your papers online makes it slightly easier to plagiarize them, but it's not really contributing much to the problem, since there are so many possible sources. (There has even been a paper plagiarized from an advertisement for a consulting firm! See http://www.siam.org/journals/plagiary/index.php. It's amazing how much garbage there is out there, and how unscrupulous some authors are.)
Instead, making your papers as visible and accessible as possible makes it more likely that plagiarists will be caught. And, of course, this is in addition to all the other benefits of making it easy to read your papers.
Vanity googling can also help: I periodically do searches for topics I care about, or to see who is citing my papers, and I've caught several plagiarists that way.
The arXiv flags papers whose text overlaps nontrivially with a previous paper. This means in fields where arXiv use is widespread, plagiarists can't get away with using the arXiv, so they are further marginalized.
Unfortunately, after plagiarism is discovered, there is little or nothing you can do to ensure that the plagiarist is punished. Reputable conferences or journals will investigate the situation and flag or withdraw the paper; less reputable ones will ignore you and hope you quit asking about it, although they will sometimes act if you publicize it enough. The plagiarist's employer probably won't do anything, no matter how serious the case is. What I'd recommend is that you try to correct the literature, and report the incident to the plagiarist's employer (assuming the paper was published as part of their job), but not worry too much about getting any action from the employer. If pursuing the case further would give you some satisfaction, then that's a good reason to do it, but don't do it with the expectation of concrete results. If you're lucky, you'll get a letter explaining that something awkward might or might not have happened, that everyone involved is very sorry for any hurt feelings, that whatever did happen wasn't really anyone's fault, but that it won't happen again. (Sadly, it often does.)
