digital trends Google just shrugged off the largest DDoS attack ever

[bonami2]

Quote

Google just reported that there was a massive attempt at shutting down one of its Cloud Armor customers, peaking at 46 million requests per second. That makes it the largest Layer 7 distributed denial of service attack ever reported, and Google’s infrastructure handled the surge in traffic.

Google explained that at its peak, this attack was equivalent to an entire day’s worth of Wikipedia visits coming in just 10 seconds. Wikipedia’s popularity places it in the top 10 of all websites, so this is an incredible feat of stability to withstand such a powerful DDoS attack.

https://www.digitaltrends.com/computing/google-cloud-blocks-largest-ddos-attack-ever/

 

Edited August 20, 2022 by bonami2

[ENTERPRISE]

A pat of on the back of the network infrastructure team! 

 

These attacks are always going to grow in strength as time moves one, but so do the defences. That being said these attacks are non AI, imagine once someone properly harnesses that one and uses it, that could be a different story. 

[bonami2]

1 hour ago, ENTERPRISE said:

A pat of on the back of the network infrastructure team! 

 

These attacks are always going to grow in strength as time moves one, but so do the defences. That being said these attacks are non AI, imagine once someone properly harnesses that one and uses it, that could be a different story. 

Never realised AI could be use for this but it true. It would fight their own AI and test it and improve while doing the attack.

[J7SC_Orion]

...very much an arms race, with AI fueling both 'attack' and 'defense' in the near future - or how about Quantum computing in tandem with AI ? ...I've got a paper map of the Yukon for a reason 

[Kaz]

On 20/08/2022 at 12:36, ENTERPRISE said:

A pat of on the back of the network infrastructure team! 

 

These attacks are always going to grow in strength as time moves one, but so do the defences. That being said these attacks are non AI, imagine once someone properly harnesses that one and uses it, that could be a different story. 

In my security + class they talked about how a company setup a large (many servers) honey pot that was infected incredibly fast, I think within 12 seconds the entire system had been compromised.  Within 24 hours the servers were compromised hundreds of time.  It was used as a warning about connecting unpatched systems to the internet.  A well designed web crawler doesn't need AI to infect things.

 

I would love to know how they successfully defend a DDOS attack while still allowing standard users to connect.  How do they identify attacker traffic from regular traffic?  It's very likely Google is already using AI to help do that.  What is going to be really scary is when AI blows a hole in all these hardware back doors.  Tech debt is a real problem and bot nets are only getting larger.  Hardware backdoors are an incredibly short sighted idea.  This is an old video, but still an awesome watch about how they have found back doors that were intentionally baked into the CPU.   The guy works for intel now, go figure.