Whether or not paradox is an issue depends on the nature of the universe. If the multiverse theory (hinted at my some aspects of quantum physics but unproven) is correct, each and every time a sub-atomic particle has a possible state change, it takes every possible path, creating a new universe as it does so. (in quantum theory, there are not pre-determined events, merely probabilities, for quanta under various conditions.) This results in an infinite number of parallel universes. Some nearly identical, others exceedingly different. Based upon that theory, a time traveler would not face paradox; he would merely be shifting to a different universe at a prior state. He could not create paradox by killing his own grandfather, as his "real" grandfather exists only in his originating universe.
Even in a linear monouniverse (currently accepted theory), it isn't clear whether the laws of physics allow, or prohibit, the existence of closed time-like curves (which is what a time machine would be creating). I'll note that Closed Timelike Curves (which by their nature violate causality) do often appear amongst the solutions to Einstein's general relativity field equations.
My best bet, based on both some experimental results in the field of quantum physics, and Einstein's general relativity, is that closed timelike curves can occur, and thus causality is not inviolate. If so, time travel is a theoretical possibility.
And BTW, if Causality is violable, then Paradox is not a problem.
I like that. It fits!
That's an interesting point about the internet. By it's nature, it's two-way communication. You need to be able to send a page request to view a page. That's why downlink-only internet does not exist (unlike a TV signal, which is true one-way communication)