CNN: U.S. has a Program to Hack North Korea Missiles. This is called “Left of Launch”

I wish this had been a “black program.” But “Left of Launch” has been peeking from behind the curtain since at least 2015. Quoting the Missile Defense Advocacy Alliance, (3/15/2015),

The “Left of Launch” strategy has been percolating over the past few years in the bowels of the Pentagon and government labs as part of an effort to reduce the cost of engagement of missile defense and to defeat an outnumbering force of offensive ballistic missiles that continue to proliferate around the world should they be used to threaten the United States and its allies.

Since CNN has given this a certain exposure, I feel I can add a little without helping North Korea. Left of Launch is practical against powers that cannot manufacture integrated circuits of a certain complexity. China, already self-sufficient or nearly so, is not in this class.

Back in the day, integrated circuits were relatively simple. A printed circuit board was required for just part of a small computer. For the most part, each chip performed a function so simple, it could be described by a “truth table”, that could be written down like a spreadsheet. You could test each chip and know exactly what it does.

The progress of integrated circuitry was so astonishingly rapid that it outran the words used to describe it. “LSI”, large scale integration, had at least a thousand gates. VLSI had many more. But some modern chips have more than a billion gates!

If you have something that complicated, and it has only a thousand pins (connections to the outside world) or so, can it be tested? Can the maker ever know whether it has any, shall we say, behavior, that is not in the specifications?

M.G. Karpovsky, a mathematician and digital logic  innovator, answered the question in the early 80’s. You can test about 71% of it. The rest is forever hidden. It would take something like the age of the universe to test all the combinations. Instead of being sure, it goes into your smartphone or laptop. At least it won’t catch fire like the battery.

A North Korean missile designer has two options. He can design the electronics for his rocket the really, really old way, using thousands of SSI (small scale integration) chips. It’s very tough to do this., and even harder to do it compactly. It was last done well by the famous names of U.S. computing. The rest gave up. Examination of scrapped Soviet electronics suggests their missiles might have not actually worked, so bad was their attempt. (Soviet missiles had a backup mechanical switch  so that if the missile actually hit the ground, it would at least not be a total dud.)

The modern alternative is to build the missile electronics out of Big Chips, with spiffy acronyms like ASIC, FPGA, and CPLD. All these chips are subject to the 71% rule. In the unknown 29%, these chips could be having their own conversation on the side. And Karpovsky has shown there is no way to test the truthfulness of a big chip.

Could Kim & Co. examine the chips with a microscope? It is doubtful they could learn anything, since an elaborate scanning system would be required for comparison with a “true” chip. Could Kim & Co. seek alternate suppliers? This is doubtful, since the technology remains in the hands of the G7. But as the technology diffuses, it could happen in the future.