ABSTRACT1I/'Oumuamua is the first known interstellar small body (Bacci et al. 2017), probably being only about 100 m in size. Against expectations based on comets, 'Oumuamua does not show any activity and has a very elongated figure (Meech et al. 2017), and also exhibits undamped rotational tumbling . In contrast, 'Oumuamua's trajectory indicates that it was moving with the local stars, as expected from a low-velocity ejection from a relatively nearby system (Mamajek 2017). Here I assume that 'Oumuamua is typical of 100-m interstellar objects, and speculate on its origins. I find that giant planets are relatively inefficient at ejecting small bodies from inner solar systems of main-sequence stars, and that binary systems offer a much better opportunity for ejections of non-volatile bodies. I also conclude that 'Oumuamua is not a member of a collisional population, which could explain its dramatic difference from small asteroids. I observe that 100-m small bodies are expected to carry little mass in realistic collisional populations, and that occasional events when whole planets are disrupted in catastrophic encounters may dominate interstellar population of 100-m fragments. Unlike the Sun or Jupiter, red dwarf stars are very dense and are capable of thoroughly tidally disrupting terrestrial planets. I conclude that the origin of 'Oumuamua as a fragment from a planet that was tidally disrupted and then ejected by a dense member of a binary system could explain its peculiarities.