So, you’ve got 100 requests per second typically. Each requests must be completed within 5 seconds, or the connection between requester and server fails. Normally not a problem, each one only takes 10 milliseconds. So 100/second x 10 milliseconds = 1 second. All good! Even if you get 5 times the normal server load, you’ll scrape by because you’ve got a 5 second timeout limit.
Now let’s say I decide to DDOS you. Im going to send you 10000 requests/second. You’re server is fucked. Ypu’ll respond to only 1/100 of MY requests, meaning the odds you respond to one of your customers is almost nil.
Thats the basic premise. Many servers have automatic halts to protect themselves, meaning if they get an influx higher than either a set number or a multiple of the last batch, it just scrubs requests for a while then reboots. It also usually happens in the millions of requests, with the DDOS being billions
So, you’ve got 100 requests per second typically. Each requests must be completed within 5 seconds, or the connection between requester and server fails. Normally not a problem, each one only takes 10 milliseconds. So 100/second x 10 milliseconds = 1 second. All good! Even if you get 5 times the normal server load, you’ll scrape by because you’ve got a 5 second timeout limit.
Now let’s say I decide to DDOS you. Im going to send you 10000 requests/second. You’re server is fucked. Ypu’ll respond to only 1/100 of MY requests, meaning the odds you respond to one of your customers is almost nil.
Thats the basic premise. Many servers have automatic halts to protect themselves, meaning if they get an influx higher than either a set number or a multiple of the last batch, it just scrubs requests for a while then reboots. It also usually happens in the millions of requests, with the DDOS being billions