A Torah scroll is a very long piece of parchement with two poles ("atzei chayim") at either end, while a Megillat Esther is a not-as-long piece of parchment with one pole at the end of the book.
Why the difference? Two poles seems the most convienient, as it allows you to save your place when you close it. Why not do that for the Megillah as well?
(Note that not every m'gila has a pole at one end. Some do, though, as you note. See Mishna B'rura 691:16.)
Aruch Hashulchan (Orach Chayim 691:7) explains:
A sefer Tora (Torah scroll) requires two poles because we read from it constantly; for n'viim and k'suvim one pole suffices.
I suspect the intent is that a sefer Tora needs the greater stability that the poles provide because it's used so often; see also the comments on this answer. (AHS notes further (there) that one possible reason for a m'gila to have no pole at all (rather than one) is that it's used so very infrequently.)
It seems that Torah scrolls didn't always have 2 poles but that it was added for greater maneuverability. See this Hakirah article (p.210 - 211) for some speculation on when and why this change was made.
Haran suggests the bars were adopted from Rome. The second bar was added to avoid having to unwind and rewind the scroll at every use, and was used in Rome by at least the first century CE.
Based on literary sources, Haran argues that a single bars was used for all scrolls until near the end of the Talmudic era, when double bars became standard. There is no recorded reason for the change and it is likely that it was done for pragmatic reasons.
See "Torah and Bible Scrolls in the First Centuries of the Christian Era", Shnaton 10:100-101. (In Hebrew.)
