KwaZulu involved a great deal of reorientation of society as a whole via the intensification of the system of amabutho (age-regiments) and the creation of the amakhanda (regimental towns), where people from all over the kingdom might serve many years minding cattle and, for female amabutho if held, raising crops and doing other sorts of production. (Ironworking, I'm not sure about--that was a male occupation, and amakhanda could make and repair objects, but normally the men would be 'junior' and not of status to be ironworkers--not sure.) That said, early KwaZulu was absolutely a conquest state and ultimately not sustainable in the form it took under Shaka and Dingane. The mode of warfare predominant in the region was cattle raiding, but Zulu warfare involved the destruction of societies and leaderships and incorporation of people as clients by force, regardless of their ongoing resistance. This also involved the total acquisition of their cattle and grain, and a transfer of that wealth and sustenance to the Zulu hierarchy at a time when the land was still suffering from dislocations of drought and war. The problem that emerged in the late 1820s and early 1830s was that societies increasingly adapted in one way or another to Zulu warfare via defensive alterations and alliances with new forces (Barolong), restructuring of societies (Lesotho and Swaziland), or simple centralization of military power (Bopedi). The heyday of raiding came to a certain end with the final destruction of Dingane's power, already blunted by his defeat in 1838 at the hands of Boer trek parties, by a combined force of those Natal Boers and his brother (and successor) Mpande's own loyal amabutho.

The restructuring of the state into something that marshaled resources more lightly, but remained markedly militarized, was not an easy one because Zulu identity was (per Michael Mahoney's The Other Zulu) not widely carried as a primary one around 1840--indeed, KwaZulu was an ephemeral thing beyond its conquest capacities. Indeed, many people left KwaZulu during the rule of Mpande and that of Cetshwayo because colonial Natal was, despite its weakness, not under any threat of Zulu power. But at times when the Zulu state was not fighting wars, the amabutho operated as training schools for the nation, a force-in-being that could be made into amampi (plural of impi or a force of variable size) for whatever purpose. Militarization and civilianness, if that's a word, were not so separate, and in an ideal Zulu form male homesteaders would be retired members of amabutho, married possibly to a woman from a female ibutho, with a few cattle from the state to build a homestead for production of families. However, we don't have a good sense of just how productive the regimental town (ikhanda) was in terms of food, only that it was supposed to make a surplus when an ibutho was in residence. Aside from internal conflicts, war was not a constant disruption after Dingane, and harvests did not suffer as often as under the worse conditions of the 1810s--so the state survived long enough to see a renaissance of Zulu identity later on in opposition to the British after the 1879 war but especially 1887 (imperial annexation) and other groups of people under colonial rule following Natal's takeover of administration in 1897.

For structures of KwaZulu, Laband's Rope of Sand (1998) is important here, although John Wright has a few essays that talk about the general features at some length. It is not as neat as Laband portrays it, though he tells a fantastic story, and Mahoney (mentioned above) complicates it in interesting ways. I still have to read E. A. Eldredge's very recent Kingdoms and Chiefdoms of Southeastern Africa which is problematic for its acceptance of some accounts at face value, but she is also powerfully analytical as a writer and knows a lot of otherwise very obscure accounts that are invaluable to anyone looking at the early Zulu kingdom.

[edit: I should point out that evidence suggests that Zulu forward forces lived off the spoils in the early years--traveling more lightly with a limited supply train. That independence was part of the reason why someone like Mzilikazi of the amaKhumalo (Ndebele) could break away while on campaign (1820s) and try to build his own conquest kingdom somewhere.]