From the beginning of time, whenever an individual or the leadership of a group has perceived the benefits of using violence as outweighing the costs, they have utilized violence. This will continue till the end of time, almost by definition. Organizations such as the UN are supposed to raise the costs of violence but that basic premise still holds. Therefore, I consider the distinctions between offensive, defensive, preemptive, and preventive war a distraction to the real argument.
If all we're trying to prevent is a few thousand or even a few tens of thousands of U.S. citizens murdered and a few big buildings knocked down every few years, then the cost of invading Iraq may not have been worth the benefits.
But what about the value of civilization itself? It's not inconceivable that a few well placed nukes could so shake the faith people have in civilizations and its institutions that the whole thing collapses like a house of cards. For example, currency only has value because: a) people think it does and b) people think that the issuing government will be around essentially forever. The latter could possibly be called into question with an attack not all that much bigger than 9/11. Many of civilization's institutions are like that so the whole thing could unravel fairly quickly.
If civilization collapses, then what? The planet would probably struggle to support even one billion people living without the structure and efficiencies of civilization. That means at least five billion dead, of all races and ethnicities. If you estimate the probability at one in ten-thousand, that's still an expected value of 500,000 dead. Even at one in a million, that's still 5,000 dead. That's how I look at the utility analysis. Thus I'm pretty easy to convince to give war a chance.