The era of nonproliferation is over. During the first half-century of the nuclear age, safety lay in restricting the weaponry to major powers and keeping it out of the hands of rogue states. This strategy was inevitably going to break down. The inevitable has arrived.

The six-party talks on North Korea have failed miserably. They did not prevent Pyongyang from testing a nuclear weapon and entering the club. North Korea has broken yet again its agreement to reveal all its nuclear facilities.

The other test case was Iran. The EU-3 negotiations (Britain, France and Germany) went nowhere. Each U.N. Security Council resolution enacting what passed for sanctions was more useless than the last. Uranium enrichment continues.

When Iran's latest announcement that it was tripling its number of centrifuges to 9,000 elicited no discernible response from the Bush administration, the game was over. Everyone says Iran must be prevented from going nuclear. No one will bell the cat.

The "international community" is prepared to do nothing of consequence to halt nuclear proliferation. Which is why we must face reality and begin thinking how we live with the unthinkable.

There are four ways to deal with rogue states going nuclear: pre-emption, deterrence, missile defense and regime change.

Pre-emption works but, as a remedy, it is spent. Iraq was defanged by the 1981 Israeli airstrike, by the 1991 Gulf War (which uncovered Saddam's clandestine nuclear programs) and finally by the 2003 invasion, which ended the Saddam dynasty.

A collateral effect of the Iraq war was Libya's nuclear disarmament. Seeing Saddam's fate, Moammar Gaddafi declared and dismantled his nuclear program. And if November's National Intelligence Estimate is to be believed, the Iraq invasion even induced Iran to temporarily suspend weaponization and enrichment.

But the cost of pre-emption is simply too high. No one is going to renew the Korean War with an attack on Pyongyang. And the prospects of an attack on Iran's facilities are now vanishingly small. What to do?

Deterrence. It worked in the two-player Cold War. Will it work against multiple rogues? It seems quite suitable for North Korea, whose regime, far from being suicidal, is obsessed with survival.

Iran is a different proposition. With its current millenarian leadership, deterrence is indeed a feeble gamble, as I wrote in 2006 in making the case for considering pre-emption. But if pre-emption is off the table, deterrence is all you've got. Our task is to make deterrence in this context less feeble.

Two ways: Begin by making the retaliatory threat in response to Iranian nuclear aggression so unmistakable and so overwhelming that the non-millenarians in leadership would stay the hand or even remove those taking their country to the point of extinction.

But there is an adjunct to deterrence: missile defense. Against a huge Soviet arsenal, this was useless. Against small powers with small arsenals, i.e., North Korea and Iran, it becomes extremely effective in conjunction with deterrence.

For the sake of argument, imagine a two-layered anti-missile system in which each layer is imperfect, with, say, a 90 percent shoot-down accuracy. That means one in 100 missiles gets through both layers. That infinitely strengthens deterrence by radically degrading the possibility of a successful first strike. Even Mahmoud Ahmadinejad, sitting on an arsenal of, say, 20 nukes, might recoil from these odds — given the 100 percent chance a retaliatory counterattack of hundreds of Israeli (and/or American) nukes would make Iran a memory.

Of course, one can get around missile defense by using terrorists. But anything short of a hermetically secret, perfectly executed, multiple-site attack would cause terrible, but not existential, destruction. The retaliatory destruction, on the other hand, would be existential.

We are, of course, dealing here with probabilities. Total safety comes only from regime change. During the Cold War, we worried about Soviet nukes, but never French or British nukes. Weapons don't kill people; people kill people. Regime change will surely come to both North Korea and Iran. That is the ultimate salvation.

But between now and then lies danger. How to safely navigate the interval? Deterrence plus missile defense renders a first strike so unlikely to succeed and yet so certain to bring on self-destruction that it might — just might — get us through from the day the rogues go nuclear to the day they are deposed.

We have entered the post-nonproliferation age. It's time to take our heads out of the sand and deal with it.

Charles Krauthammer is part of The Washington Post Writers Group. Reach him at letters@charleskrauthammer.com.