I only made this to feed my own curiosity (and for a classroom homework too tbh) since it's not very effective nowadays, but feel free to use it! Python 3 required (if you want to use python2 change the print near the end of the script)

To use it run

pip install -r requirements.txt 

to install the required dependencies

Usage: inventedAttack.py [OPTIONS]

Options:
  -i, --ip TEXT          IP address of the target machine
  -p, --port INTEGER     Port of the service to attack with SYN Flood
  -t, --threads INTEGER  Number of concurrent threads
  -s, --size INTEGER     Fragment size
  --help                 Show this message and exit.

If you don't pass any of the parameters, the script will ask for them with an interactive prompt

I actually tried it on some machines and it had 0 impact, probably because the base of the attack (SYN Flood) was effective when resources were scarcer and the was no SYN Flood protection builtin in the kernel. If you want to play around the idea anyways take a look at

/proc/sys/net/ipv4/tcp_syncookies

/proc/sys/net/ipv4/tcp_max_syn_backlog

/proc/sys/net/ipv4/tcp_synack_retries

If you want to monitor the half-open connections on the server you can try

netstat -tuna | grep :443 | grep SYN_RECV

Change 443 for whatever port you are using, also you can pipe again | wc -l to count the number of connections made, in my tests they stay in the range of 50-100.

The fragmentation part makes no difference either.

About the source IP spoofing I found it was the most effective part since the web server was making DNS PTR requests for each random source IP, so it kind flooded the DNS with them.

Lessons learned : turn off reverse DNS resolution in your services.

Things I might improve:

• Write this in Python 3 (I actually don't know why I was sing python 2, the only incompatible function was a print!)
• Use Python 3 async, see how performance improves Asyncio makes no difference at all, but I'm leaving anyways a branch here https://github.com/avantasia/inventedAttack/tree/async for testing purposes (and to remind me the next time)
• More configurable parameters (fragment size, threads)
• Tests with WAF/IDS and fragmentation on target machine
• Performance graphs on target machine
• Maybe dockerize both this and target machine

David Carracedo Martinez - [email protected] 2019