Hi,
I want to ask a question about IP spoofing.
I read some papers on it thoroughly, like the papers on IP spoofing on SANS and Phrack magazine (for the links, see lower).

These papers are quite aged (1996, 2000) and tell me about the 64K rule and time-based sequence number increasings. I suppose that, because this TCP vulnerability are widely known for a long time, new operating systems have accurate prevention mechanisms (like better random numbers generators and other things)

Now my question is, in what extent are IP spoofing attacks possible in the current operating systems? What has changed and áre the attacks still possible?

can someone answer this?

greets,
prinski

the papers i read:
* IP Spoofing Explained - From Phrack Magaizine by daemon9 / route / infinity
* A short overview of IP spoofing: PART I by Brecht Claerhout
* A short overview of IP spoofing: PART II by Brecht Claerhout
* Very simple explanation of IP spoofing by Victor Velasco

I don't think your links have shown up because you don't have enough posts to post a link yet...

But yes it is still possible. They are not easy, which I'm sure the article would've pointed out. I look throught my firewall logs at both home and work and see it all the time where people have tried to get a response by using a spoofed "internal IP address"

Well it depends how completely you want to spoof. The UDP transport, for instance, has not "connection" state, so there are no sequence numbers to maintain. For that reason anyone can send a UDP datagram with a spoofed source--same with ICMP. For TCP there are sequence numbers, but as mentioned above some times they're predictable. Many operating systems have taken steps to correct that, so for instance the *BSD family, Linux, and Windows have all changed their TCP sequence generation, and even changed the way they fill some unused flags in the TCP headers.

What this doesn't eliminate is a MitM (Man in the Middle) attack. If you're sitting on a network device between the two parties on a TCP connection, you can intercept the datagrams from each end and forward specially crafted datagrams in their place. Neither end would be the wiser, since the sequence numbers and checksums would agree.

Various countermeasures have been designed to prevent MitM attacks, such as SSL/TLS, IPsec, and IPv6.