The stocky man standing before me was immaculately turned out in a dark blue pin striped suit. With his thick New Jersey accent he could have been a movie Mafioso. But he wasn’t. Despite the cognitive dissonance this situation wasn’t as funny as it seemed. This apparent movie gangster was briefing me on Armageddon: full-scale nuclear war. He talked about a five-minute war – where all the nuclear weapons arrived at their targets simultaneously. He talked about a twenty-minute war: The missile launches would be simultaneous so that different targets, at different distances, would receive their doom at different times. He talked about megadeaths. He talked about the forever future of the world being determined in an hour. The subject was dead serious,  for we were employed in the business of deterring such a catastrophe.

Nuclear weapons have three essential characteristics: They are very expensive, they must be delivered, and they are fearsome. These aspects dominate all modern strategic thinking.

Consider, first, the cost. Producing a fission bomb is a very expensive proposition. The old rule of thumb was $100 million for a regular production fission device. A hydrogen bomb is much more difficult and expensive. Developing just the capability to make such bombs is vastly more expensive than the production bombs, themselves. The real numbers are unknown except to a few. Moreover, making such devices small enough, compact enough, and lightweight enough to be useful as weapons is a nontrivial exercise.

Everything considered, the cost of these weapons is a stretch even for a well-developed economy. For a marginal economy, the cost of autonomous development is a back-breaker. It is usually cheaper to buy these things if they are available.

Because of their high cost, nations are economically inhibited from actually using nuclear weapons. They are usually considered both a prestige item and a deterrent. India and Pakistan both have long had deliverable nuclear weapons. Neither nation has been inclined to use them even though they have occasionally been at war with each other.

In the past, nations that have nuclear weapons have acted rationally rather than suicidally. But not all nations are rational. North Korea plainly is not. And, too, Iran has leaders who await the Twelfth Imam -- the Mahdi -- and the end of the world.

Having a bomb is not particularly useful unless it can be delivered. There are three existing methods of delivery: surface, airborne, and ballistic missile.

Surface delivery is by boat, truck, or cargo container. Existing radiation sensors can detect many types of bombs, but only at close range -- a matter of yards. Thus, such weapons can be difficult to detect. Bombs must be funneled past sensors in order to be detected. We do that now at several ports of entry. Small boats and disbursed trucks are much more challenging. Only the future will tell if this kind of smuggling can be stopped. In any case, surface delivery can only wound a continental nation, not kill it. Thus, surface delivery is only useful for terrorism or blackmail.

Airborne delivery has old, and well-established, solutions. Effective bomber defense was developed in the 1950s.

Ballistic missile delivery is the current challenge. Long range ballistic missiles have three flight regimes: boost phase, exoatmospheric, endoatmospheric.

The best way to kill a missile, and its warheads, is in its boost phase when the missile is most vulnerable and its fiery rocket engines keep it from hiding. But boost phase interception requires that the defensive weapon be in a position to intercept the missile. This usually means space basing. Earth orbiting space-based High Energy Lasers can reach out over thousands of kilometers. So mere dozens of HEL battle stations can do the job. Space-based interceptor rockets, on the other hand, are constrained by their velocities. For the boost phase defense, up to thousands of space-based interceptor rockets may be needed.

Airborne lasers can kill up to hundreds of kilometers, but they must patrol outside the hostile’s borders – and therefore can only reach a limited distance into his territory. If one is willing to violate an adversary’s territory, then interceptor rockets could be mounted on high-flying stealth drone aircraft so as to circle over potential launch sites.

Exoatmospheric interception is probably the toughest system level challenge. This is not because it is hard. Rather, it is because of the geographical dynamics of the situation. The interceptors and sensors must be properly sited. The sensors must be close enough to the flight path see what is happening despite the Earth’s curvature. The interceptors must be able to reach the deployed warheads.

In this respect, it should be noted that President Obama’s abandonment of sensors and interceptors in the Czech Republic and Poland was pure appeasement of Russia and pure betrayal of Europe. The withdrawal made no technical sense. Such interceptors would work against an Iranian attack on Europe or the U.S. But they could not intercept Russian missiles unless Russia was attacking Europe. The trajectory dynamics precluded intercepting Russian ICBMs aimed at the U.S.

More critically, the missile defense radars, and other sensors, must be able to see and track the missiles through boost, and the warheads throughout midcourse. Picking out the warheads while they are coasting through space accompanied by sophisticated decoys is a technical challenge. But, indications are that it is solvable, or has been solved. Once a warhead has been located, killing it is not all that difficult. Hitting a bullet with a bullet is actually rather easy.

What makes this possible? The answer is the self-adapting servo. The thermostat which controls the heating or cooling of your house is a simple servo controller. A servo is has a sensor which detects the error when a desired condition is no longer met. It also has an actuator which operates in such a way as to reduce, or eliminate, the error.

The anti-missile interceptor has a camera as well as a computer which tells the interceptor the direction to the target. This is the sensor. It also has various thrusters which guide the interceptor to the target. This is the actuator. The guidance algorithm is very simple: just pulse the various thrusters so as to keep the image of the target centered. Simple, yet effective.

Endoatmospheric defense is the easiest to do from a technical standpoint. Except that it also requires the greatest resources. That is a matter of geography. Once a warhead enters the atmosphere its deceiving decoys are stripped away, leaving the warhead nakedly visible. However, to intercept the warhead, the interceptor must be stationed near the warhead’s target. Since it is impossible to know ahead of time the intent of the enemy, interceptors must surround each and every high-value target. This means a massive investment in interceptor systems because a modern nation has so very many high value targets.

The most promising, and economical, missile defense is boost phase interception by Space Based Lasers. This appears to be the best way to defend a nation against missile attack. The reason is that a relatively small number of SBL battle stations can cover the globe and protect against a missile launch from any location to any destination. (Full disclosure: I have many years of experience working professionally on SBL systems.)

Some form of “layered” defense where all these defensive elements are in play makes up the most desirable architecture. But resistance to the space-based part of the system is very strong.

The real problem with SBL’s is political -- not technical. Those who wish to continue the policy of Mutually Assured Destruction (MAD) lobby strenuously against developing an SBL system. They argue that such a system crosses the line against putting weapons in space. Of course, the warhead from a ballistic missile is already very much a weapon in space. Its launch vehicle is simply stored on the ground. So the weapons-in-space threshold has long since been crossed. The real objection to an SBL system is that it promises to be too effective. MAD’s foreign policy constituency would be out of a job.

I see no moral impediment to basing a purely defensive weapon in orbit. This argument will ultimately be resolved when someone detonates a nuclear weapon in anger, as will likely happen.

There is only one correct way to look at missile defense. That way ignores its monetary cost. That way ignores the proportionality cost trades of offense versus defense. That way focuses only the consequences of not having missile defense: Missile defense is insurance. Why buy this insurance? If a nuclear weapon is exploded in downtown Manhattan is the resulting cost less than, or greater than, the cost of an effective, comprehensive missile defense?