A male spouse, one half of a pair. Looking at the word, I expect it to be derived thus: hus=house and band=bind, bound. The OED definition, not too surprisingly, is all about this:
Late OE; húsbonda, -bunda, f. hús house + late OE. ? bónda, bonda, bunda, a. ONor. bóndi, peasant owning his own house and land, freeholder, franklin, yeoman; earlier búandi, bóandi, orig. pres. pple. of búa, bóa to dwell, have a household; but the OE. use answered immediately to ONor. húsbóndi, a man of this rank in his capacity as head or master of the household. In ME. often with connective e, as in husewif, housewife.
House-bound, bound to the house. This seems to me to refer to the medieval concepts of the hold, the copyhold, held land and the household Very Strongly Indeed. The husband is bound to the house as well as to the wife (i.e. housewife, huswife--the word "housewife" is very different in derivation from the 1950s ideal, thank god). They are bound to the land; they husband it as well as the house. The house or "hus" in itself stands for every part of the hold--the householding does not only include the people sustained by the land, but the land itself, the house, everything. Everything is held under the husband, the householder, he who is bound to the house and thus to the household.
The verb "to husband" as well is interesting with this land business. A husbandman keeps the land, or husbands it. He takes physical care of it, and thus takes physical care of the household as a whole: he grows crops or breeds animals. These then sustain the household. Nowadays a husband would take more than physical care of everything, and I would think that he would have as well in the Middle Ages: one cannot really be filled with well-being if one is starving, for instance. The husband provides. What he specifically provides is different in each individual case. However, by definition, he does provide.
Addendum, 11 Dec.: Gritchka tells me that "-band" is a German participle, with root "to be" such that it equals "be-ing". This is also pretty interesting. Is the husband then a being of the house? One who lives or is of the house? This seems to match with my idea of binding pretty well; if one is a being of x, then one is part of x. The husband is bound to the house even if the derivation might not state this explicitly. So I think that the idea of binding is still relevant here.