Schneier on Security

Roger • May 24, 2006 8:56 PM

how to squid-proof all the computer systems

8^) The answer is Astroturf. For some reason, cephalopods can’t crawl over it or grip it. So aquaria suffering from excessive cephalopod somnambulence line the lip of the tank with Astroturf. So, more furry green consoles!

Back to topic: the article as presented does really seem quite dumb. But seen in the context of the paragraph from which it was linked it isn’t quite so obviously silly:
“The bill’s goal was to ensure that the Pentagon had secure suppliers, but had it become law, it also would have had the effect of controlling bogus parts [see sidebar, “Counterfeit Electronics as Weapons of Mass Disruption?”]. ”

In other words, the claim is not that terrorists might use bogus electronics to attack consumers, but that foreign companies controlled by hostile governments might engineer well-hidden trapdoors for nefarious purposes (espionage being perhaps more likely than sabotage).

This is not so improbable because several such examples are known to have actually occurred. There are obvious countermeasures. For example, you could reverse-engineer every new mod of every chip supplied, and x-ray, say, 10% of all received chips to ensure they do not deviate from the one you examined. However this would be far more expensive than using a trustworthy supplier, and even then you could not guarantee that the supplier could not outsmart you somehow, with features which do not show up on the x-ray, or which your reverse engineers do not understand.

Roger • May 25, 2006 8:27 PM

@Alex:

While I don’t disagree with you that the story is unlikely as given, I’d like to make a couple of minor corrections:

If the French were somehow interfering with the software (how? – not networked)

French technicians were working in Argentina from November 1981 till the end of the war, outfitting and servicing the missiles. The most common version of the claim is that they sabotaged the control software directly.

It has been claimed that it was done with “secret codes” — but only by Mitterand’s psychoanalyst, who is not regarded as an expert in missile control. Notably someone who does know what he was talking about, Sir John Nott (British Secretary of State for Defence during the war) has claimed in his memoires that France helped British intelligence to thwart Argentina’s attempts to source grey-market replacement Exocets. Conceivably it is a reference to this which Mitterand’s psychoanalyst overheard.

Secondly, “how? – not networked” — well, Exocets have active radar terminal guidance. That means they have a (powerful, specialised) radio transceiver. And while I don’t know it for a fact, I would be very surprised if this radar did not include EFF capabilities. I don’t think that anything like that was done because it would have been far more effective, but it certainly is not technically impossible.

… MV Atlantic Conveyor and 1 HMS Sheffield …

Your numbers don’t seem quite right; what about HMS Glamorgan, the third British ship hit (which was moderately damaged by fire, but did not sink because the warhead failed to explode)? Oh, I see, the attack on HMS Glamorgan was launched from a shore battery (of two missiles salvaged from the damaged ARA Guerrico’s ship-borne launchers) and you’re restricting it to “airlaunched Exocet”. OK, but for the purposes of this analysis it is obviously irrelevant whether the Exocets where ground, ship or air launched (the Argentinians eventually had all 3), unless there is an argument that the French technicians only had access to airlaunched missiles. However, even accounting for that your list doesn’t seem to accord with most accounts.

80% of them flew, and 75% hit their targets, 100% of which sank

I make that 86% of known launch attempts flew, with 6 successful launches; 50% of those (3) hit targets (in one case not the intended target); of those 67% (2) (possibly 100%) failed to explode; and of the three hits on ships one (33%) directly resulted in sinking, one (33%) indirectly resulted in sinking during a storm a week later, and one (33%) did NOT sink and was not even crippled.

The repeated failures to launch and 50% failure to hit (or 67% failure to hit intended target) could be blamed on software sabotage, although I doubt that was the case. The 67% failure to hit intended target could be attributed to very effective RN ECM, which may possibly have been assisted by the French providing detailed specifications of the guidance system. However the very high rate of failing to explode (possibly as high as 100%) appears to be an inherent flaw in the missile’s design, as it was also observed with the much larger numbers of Exocets fired in the Iran-Iraq war.

In detail, most accounts list the Exocets fired as:
* multiple, repeated failures with the launch mechanism, nearly all of which were eventually rectified and Argentina ended up firing all the airlaunched Exocets it possessed plus one shore-launched naval one (most naval ones were lost when the General Belgrano was sunk);
* 2 airlaunched Exocets fired at HMS Sheffield (mistaken for HMS Hermes) on 4th May 1982, of which one passed by HMS Yarmouth without locking on and crashed at sea, and one struck HMS Sheffield but failed to explode, causing a severe fire which resulted in HMS Sheffield sinking a week later when struck by a gale whilst under tow;
* 2 airlaunched Exocets fired at HMS Invincible on 25th May, of which one crashed at sea (possibly after being hit by antiaircraft fire), while the second was decoyed with chaff but relocked onto MV Atlantic Conveyor, hit it, may or may not have exploded (witness accounts differ), and started a severe fire which caused it to sink later the same day;
* 1 airlaunched Exocet fired at HMS Invincible on 30th May, which the Argentinians claimed as a hit but Britain reported the missile was shot down by antiaircraft fire. Certainly if it was a hit — which seems extremely unlikely — then it did very little damage and caused no loss of life. The missile attack was followed up by iron bomb and cannon attacks by Argentinian Skyhawks, which apparently actually hit the hulks of other vessels which had already been damaged by bomb attacks.
* 1 ground launched Exocet fired at HMS Glamorgan on 12th June, hit, missile failed to explode, a fire was started but brought under control. The ship returned to the UK under its own power and was returned to service after a few months. Second missile in the same battery possibly failed to launch.
* Despite the great concern for the capabilities of the Exocet and new French built Super Etendards, far more damage was done by iron bombs mostly delivered by 30 year old Skyhawks. Overall 4 other RN ships (HMS Coventry, HMS Ardent, HMS Antelope, and RFA Sir Galahad) were sunk by iron bombs, and 7 others damaged.