Hacker News Reader

Comment by Quarrelsome | Hacker News Reader

nothing like this scale. These machines are full blue screen and completely inoperable.

The problem is concentration risk and incentives. Everyone is incentivized to follow the herd and buy Crowdstrike for EDR because of sentiment and network effects. You have to check the box, you have to be able to say you're defending against this risk (Evolve Bank had no EDR, for example), and you have to be able to defend your choice. You've now concentrated operational risk in one vendor, versus multiple competing vendors and products minimizing blast radius. No one ever got fired for buying Crowdstrike previously, and you will have an uphill climb internally attempting to argue that your org shouldn't pick what the bubble considers the best control.

With that said, Microsoft could've done this with Defender just as easily, so be mindful of system diversity in your business continuity and disaster recovery plans and enterprise architecture. Heterogeneous systems can have inherent benefits.

If you have a networked hybrid heterogeneous system though now you have weakest link issue, since lateral movement can now happen after your weaker perimeter tool is breached

A threat actor able to evade EDR and moving laterally or pivoting through your env should be an assumption you’ve planned for (we do). Defense in depth, layered controls. Systems, network, identity, etc. One control should never be the difference between success and failure.

https://apnews.com/article/tech-outage-crowdstrike-microsoft...

> “This is a function of the very homogenous technology that goes into the backbone of all of our IT infrastructure,” said Gregory Falco, an assistant professor of engineering at Cornell University. “What really causes this mess is that we rely on very few companies, and everybody uses the same folks, so everyone goes down at the same time.”

WannaCry did about the same damage to be honest. To pretty much the same systems.

The irony is the NHS likely installed CrowdStrike as a direct reaction to WannaCry.

The difference is malware infection is usually random and gradual. CrowdStrike screwup is everything at once with 100% lethality.

Computers hit by ransomware are also inoperable, and ransomware is wildly prevalent.

Yes, but computers get infected by ransomware randomly; Crowdstrike infected large amount of life-critical systems worldwide over some time, and then struck them all down at the same time.

I'm not sure I agree, ransomware attacks against organizations are often targeted. They might not all happen on the same day, but it is even worse: an ongoing threat every day.

It's why it's not worse - an ongoing threat means only small amount of systems are affected at a time, and there is time to develop countermeasures. An attack on everything all at once is much more damaging, especially when it eliminates fallback options - like the hospital that can't divert their patients because every other hospital in the country is down too, and so is 911.

Ransomware that affects only individual computers died not get payouts outside of hitting extremely incompetent orgs.

If you want actually good payout, your crypto locker has to either encrypt network filesystems, or infect crucial core systems (domain controllers, database servers, the filers directly, etc).

Ransomware getting smarter about sideways movement, and proper data exfiltration etc attacks, are part of what led to proliferation of requirements for EDRs like Crowdstrike, btw

Ransomware vendors at least try to avoid causing damage to critical infrastructure, or hitting way too many systems simultaneously - it's good neither for business nor for their prospects of staying alive and free.

But that's besides the point. Point is, attacks distributed over time and space ultimately make the overall system more resilient; an attack happening everywhere at once is what kills complex systems.

> Ransomware getting smarter about sideways movement, and proper data exfiltration etc attacks, are part of what led to proliferation of requirements for EDRs like Crowdstrike, btw

To use medical analogy, this is saying that the pathogens got smarter at moving around, the immune system got put on a hair trigger, leading to a cytokine storm caused by random chance, almost killing the patient. Well, hopefully our global infrastructure won't die. The ultimate problem here isn't pathogens (ransomware), but the oversensitive immune system (EDRs).

I want to agree with the point you're making, but WannaCry, to take one example, had an impact at roughly this scale.

I think recovering from this incident will be more straightforward than WannaCry.

At large-scale, you don’t solve problems, you only replace them with smaller ones.

Reply on news.ycombinator.com