site banner

Israel-Gaza Megathread #2

This is a refreshed megathread for any posts on the conflict between (so far, and so far as I know) Hamas and the Israeli government, as well as related geopolitics. Culture War thread rules apply.

6
Jump in the discussion.

No email address required.

Pretty interesting analysis of the complex systems fuckup that the Hamas attack was:

Some former members of the IDF who served on the border have in recent days testified on social media that the fence really was a technological marvel. Not so much as a stray cat could get anywhere near the border without setting off alarms, they recall. And the government and military certainly seem to have believed it was indeed impenetrable and really had changed the reality on the ground; hence partly why, by the start of this month, they had redeployed most of their regular military forces to guard the West Bank and northern border instead.

...

In any case, Hamas was able to begin their attack with the element of surprise. This was aided by an initial early-morning barrage of rocket fire, which was a relatively routine experience for the Israeli garrison forces, but which survivors recall sent most of their number hurrying as a standard precaution into fortified bunkers where – critically – they could not physically observe the approach to the border. They would normally have instead relied on the surveillance cameras to monitor the situation. Hamas, however, used small, off-the-shelf drones rigged with mortar rounds and other explosives to attack and disable the communications towers powering the network. These drones were too small and low-flying for radar to detect, so would have had to have been spotted by eye and ear. Without the cellular data link provided by the towers, the cameras did not function, and neither did the sensors and alarm systems.

With the surveillance and communications systems down, Hamas commandos then used their now infamous paragliders to simply fly over the fence. There they faced little armed opposition. The remote-controlled machine gun emplacements, if they could even operate without wireless data, had also been destroyed by drones. Now isolated, 23 high-tech observation posts each manned by a single soldier – all of them young women – were ambushed and rapidly overwhelmed by the first attackers. Those who tried to report the attacks would have found they couldn’t easily communicate. Meanwhile Hamas used bulldozers and wire cutters to quickly level around 30 sections of the fence without resistance. All of this took only a matter of minutes.

Operational command and control of the IDF division guarding the border had been concentrated into a single centralized base close to the fence. As some 1,500 Hamas terrorists surged across the now open border, this base was quickly overrun and the senior officers there killed or captured. They likely received little-to-no warning, given pictures circulating of scores of soldiers having been shot while asleep in their barracks, many still in their underwear. The subsequent sudden absence of central leadership and breakdown in the chain of command, along with the communications problems, meant that the scope and gravity of the overall situation could not easily be pieced together or communicated to either local forces or to national-level military command. Thus in the end it took hours for leaders to fully grasp what was happening and for reinforcements from elsewhere in the country to be successfully contacted, mobilized, coordinated, and moved to the south to confront the threat.

In my tech career I've always treated monitors that have lost signal with the same priority as the monitored value exceeding some critical bounds. The idea being that without signal, the monitored value very well could be in a critical state, and loss of signal itself could be caused by some greater issue.

I have no military experience. Does the military not do something similar? For example, a lost camera feed must be treated as an attack until proven otherwise.

To support your point from an IT perspective, at a previous job, the server monitoring system malfunctioning is what tipped me off to a ransomware attack being triggered. Of course, as with anything this has to be calibrated so there aren't so many false positives that alerts or downtime is ignored, but an otherwise robust system going down for seemingly no reason should arouse suspicion.

I'd be curious to see a timeline of the entire event. Maybe they were able to time it quickly enough that the people monitoring these things didn't have a proper chance to respond before the para-gliders were on top of them.

The failure rate on, say, CCTV cameras is high enough that it’s not tenable.

Source: my manager was assigned to Afghanistan as a surveillance contractor. He noted that they had to run the cables between cameras at the top of the walls, or rats would chew them overnight. And they’d still try their best to get at the anchor points.

https://www.astralcodexten.com/p/heuristics-that-almost-always-work

The whole point of running a surveillance system is to freak out when something goes wrong, even if it's rats 99.9% of the time. We have a burglar alarm at our summer cabin and so far it has been triggered only by power loss or by one of the family members tripping it because the were deep in their thoughts. I am sure it's the same for everyone else in the area and a sign that says "This home is protected by Unity Security Services" is enough of a deterrent. But the deterrent only works if you actually know that if you trip the alarm, the patrol will be there quicker than you can get that TV over the fence, even if 99% of the time they are met by the embarrassed owner who really had to take a dump pronto.

Or take fire drills. Can you imagine how terrible it is to walk all the way from the 70th floor of your skyscraper after you've sprained your ankle jogging? You need to find that weird-looking wheelchair with grippy runners, find someone willing to push it down one hundred and thirty eight flights of stairs, it's always raining when there's a fire drill because why wouldn't it, and then you have to limp back to the elevators and explain to your boss why your report is overdue. You sprain your ankle on September 10th, 2001 and limp to work the next day.

That's a fine theory, but it goes against human nature to expect people to not detect patterns like: each time I investigate, it turns out to be a rat. So you then need a mechanism to prevent people from acting normally, which is a hard problem to solve.

it seems to me that this kind of mass attack will always succeed to some extent. maybe it was made worse in this particular situation for a bunch of reasons but even if everything went right for the israelis i think hamas would have had some kind of success. unless you have some kind of massive DMZ and large permanent deployment of troops an enemy will always be able to surge at a critical point and have some short term success.

There are different definitions of success. Palestinians breaching into Israel and having engagements with IDF is one thing. The slaughter and carnage that ensured is different.

You can make Israel bleed - it is not hard to do. Making it hurt is harder. And however we cut it - on tactical level it's win for Hamas.

This is interesting, and it's nice to have confirmation (?kinda? I'm not sure how much to trust this guy) of some suspected details, or at least someone else suspecting them, but a lot of these raise further questions than they answer.

The author loves the KISS principle, and he's not wrong, but the details he's proposing are less descriptions of a complex system falling so much as a fragile system failing unnoticed. I really hope that the IDF's tower comms were not solely 'cellular', but even if he's using that as a shorthand for a combination of cellular, microwave point-to-point, and packet radio (LoRA's cheap!) that I'd consider the bare minimum for a short-term deployment, this stuff's been deployed for close to a decade and there's really no excuse to not have physical ground links and conventional radio installations. The threat of drone-delivered explosives has been present in the public info since 2017, and commercially-available data links are notoriously fragile not just to attacks but even to stuff like nearby lightning strikes.

You can't harden these things against every possible attack, but you can have enough physically separate systems that anything breaking too many of them is an obvious attack, and at least some of these tools can treat an unnatural down state from a natural one, (and some, like flares, can be a signal only available when nothing is an option). Which sounds like ass-covering, but the counterfactual environment where you had a hundred IDF soldiers on a wall against a thousand-plus Hamas soldiers with a lot of explosives still sounds like an environment where you need to call for backup, and cellular is the obvious and simple and wrong answer there, too.

Which doesn't speak to the broader point, but leaves me concerned about how precise the rest of the analysis is.

The author loves the KISS principle, and he's not wrong, but the details he's proposing are less descriptions of a complex system falling so much as a fragile system failing unnoticed. I really hope that the IDF's tower comms were not solely 'cellular', but even if he's using that as a shorthand for a combination of cellular, microwave point-to-point, and packet radio (LoRA's cheap!) that I'd consider the bare minimum for a short-term deployment, this stuff's been deployed for close to a decade and there's really no excuse to not have physical ground links and conventional radio installations.

This is largely my thinking too. The original wireless communications setup is cost effective for hundreds of kilometers of perimeter, but centralised failure points like towers are too vulnerable. Buried cable is more expensive, but resistant to jamming or drone/rpg attacks. There needs to be much more redundancy in communications.

The other major issue is probably the majority of guards running for bunkers under the rocket attack, seemingly without means to shelter in place or to monitor the perimeter. I'd give them the benefit of the doubt and presume CCTV monitors in the bunkers, but the monitors are useless if cameras were taken out and they can't physically observe the fenceline.

Regarding KISS, I think the author is conflating multiple overlaying and interconnected barriers (as in Defense in Depth) with the systems reliance on the comms tower. He's right that there shouldn't be a single point of failure for the majority of the systems in place (which seemed to be the case). There should be redundancy and the use of multiple technologies and procedures carried out by humans. Flares, Radio, heck even periscopes from the bunkers could all have been useful as failsafes. These multiple systems should ideally function independently, and by doing so become a nightmare to overcome in an assault.