The Pentagon has been bracing for trouble on the peninsula ever since the 1953 armistice. The basic plan, in place for almost 50 years, is predicated on a North Korean invasion of the South. But in the early 1990s, that plan was expanded. Titled OPPLAN 5027, it now lays out a campaign in which, the invading North Korean Army having been destroyed, U.S. and South Korean forces invade the North and topple the regime. The Pentagon even has a contingency plan for taking out all known nuclear facilities in the North.

Kim shouldn’t lose sleep over it.

Not that he should dream of invading the South. Defeating such a move would be bloody but relatively simple, the Pentagon’s planners expect. “We will halt the North Korean Army in the passes,” Colin Powell, then chairman of the Joint Chiefs of Staff, assured the Clinton administration in 1993. “Then we will kill them.” By passes, Powell meant Kaeson and Chorwon, the only two corridors from the North broad enough to allow the movement of formations the size of North Korean divisions. Those places would be killing grounds.

But repelling an invasion has never been the problem. The real issue is Seoul, a city-sprawl of 22 million—almost half of South Korea’s entire population—within range of North Korean bombardment. The U.S. military reckons that two thirds of the North’s 11,000 artillery pieces and multiple-rocket launchers are deployed in hardened shelters just behind the DMZ, with at least half within range of Seoul, capable of pounding the South with 500,000 rounds per hour for several hours. The South Korean capital is Kim Jung Il’s hostage.

The North’s attack strategy is assumed to be nothing fancy: a massive barrage on Seoul, causing a tidal wave of refugees and (supposedly) forcing the South’s government to surrender quickly under threat of further destruction of the capital. Defending the South against such a blitzkrieg would be tough. The Pentagon’s plan has U.S. air reinforcements pouring in at six main airbases, ready to fly more than 3,000 sorties a day. All of those fields are vulnerable to strikes by Northern missiles armed with chemical warheads or cluster munitions. And the fields are all near the coast. The North’s special forces (nominally more than 100,000 strong) are known to have been practicing submarine-launched assaults, and the airbases would presumably be key objectives of such assaults.

Worse yet, the Pentagon plan has U.S. Army reinforcements all arriving through a single port, Pusan, in the far south—again a target for Northern missiles. And after disembarking, the Americans would have only two good roads up the length of the peninsula, roads that would be packed with fleeing civilians and presumably under bombardment.

The Pentagon has worked like crazy trying to improve South Korea’s defenses. There has been a huge effort, for example, to set up sensor-communications-counterfire systems so each of those North Korean artillery pieces, as soon as it fires a single shell, can be pinpointed and destroyed faster than it can be rolled back into its cave. Current plans rely on massive U.S. airpower against targets in the North: stealth aircraft, cruise missiles and precision-guided munitions (a.k.a. “smart bombs). They also include a detailed outline for action on the ground, such as placing U.S. Marines deep inside the North to threaten Pyongyang. But none of these expedients looks remotely likely to cut the butcher’s bill to an acceptable level.

And so far we’re only talking about confronting an attack from the North.

Bill Clinton and William Perry, his Defense secretary, considered going to war against Pyongyang in 1994. On May 19 of that year, Perry, along with the then Joint Chiefs chairman, John Shalikashvili, and the commander of U.S. forces in South Korea at the time, Gen. Gary Luck, briefed the president on the anticipated costs of such a war: roughly 52,000 U.S. military killed or wounded; 490,000 South Korean military ditto, and untold numbers of Northern dead and civilian casualties, all in the first 90 days of conflict—together with a U.S. price tag of more than $61 billion. Luck later calculated the ultimate toll at more than 1 million dead, possibly including as many as 100,000 Americans, and a final bill to U.S. taxpayers in excess of $100 billion—not to mention more than $1 trillion in damage to South Korea’s economy.

Which helps explain why Clinton opted instead to make a deal with Pyongyang, despite the deficiencies of the 1994 Framework Agreement. “We all knew it wasn’t a solution,” says one senior Defense Department official involved in those discussions. “But it staved off war. And it bought time.” The 1994 casualty estimates “haven’t changed,” says a senior Pentagon official. “A major war on the Korean peninsula would have a disastrous effect on the region’s economies,” Maj. [now Lt. Col.] Daniel Orcutt warned in a study published by the U.S. Air Force Academy in 2004. In particular, he wrote: “War would obliterate South Korea’s goal of becoming a top-ten economic power.”

Even so, the prospect of North Korea getting the bomb worried the Pentagon sufficiently that in 1994 it launched another study, this time asking whether a way could be found to destroy the North’s known nuclear facilities without triggering all-out war on the peninsula. The Americans could obliterate the targets in short order, using long-range bombers and carrier-based aircraft to avoid implicating the South. The main thing, as described by one of those involved in drawing up the scenario, would be to make clear to the North’s leadership that the attacks were intended only to eliminate Pyongyang’s nuclear capability, not to overthrow the regime. But the Northerners would also be put on notice: if they responded by attacking the South, the regime would become a target. The scheme—it never got as far as a plan—was scrapped. Pyongyang’s reaction was simply too hard to predict.

A few things have changed since the mid-1990s. In those days America had more ground troops in South Korea than it has now, and the United States wasn’t fighting even one war back then, let alone a pair of them. Today the Pentagon could muster just about enough air and sea power to defend South Korea if the North invaded, according to the Joint Chiefs of Staff’s latest annual “apportioned forces” estimate. But doing the job would take pulling every Marine out of Iraq without delay and sending them all straight to Korea.

In May of this year, Bush signed a directive assigning the U.S. Strategic Command the task of overseeing and coordinating all U.S. plans to counter threats of WMD proliferation worldwide. Bad idea. As overseer of America’s nuclear forces, STRATCOM at once set about doing what it knows how to do, which is to plan for nuclear strikes. Duly it hatched a concept—notional, global, and nowhere near as detailed as a plan—for nuclear strikes against proliferators who are judged to pose an impending threat to the United States. The proposal inevitably leaked and was quickly quashed. Now everyone has gone back to the drawing board—except Kim’s scientists. They’re doing just fine.