IP Spoofing

From Computing and Software Wiki

Revision as of 02:59, 13 April 2009 by Luongp (Talk)
Jump to: navigation, search

IP(Internet Protocol) Spoofing is term used to describe the creation of IP Packets with a forged

(spoofed) source IP address for the purposes of hiding the true identity of the sender or impersonating

the identity of another system. IP Spoofing is generally used to gain unauthorized access to a network by

impersonating a source with authorized access. IP Spoofing attacks are based on the fact that internet

communication occurs between distant computers. The communication is handled by routers which decide the

best path based on the destination address while the source address is generally ignored. Once the

destination address has been reached, the destination machine reacts based on the source address.

Contents

How IP Spoofing Works

Information is transferred over the internet in the form of Internet Protocol(IP) Datagrams also known as

Packets. IP Datagrams consists of the IP header and the data being transferred shown in Figure 1. It is

the IP header that contains the IP Header contains the source IP address and IP destination address.

Image:Ipheader.jpg

Normally IP Packets are created automatically behind the scenes for the user; however, a user can override

this function by inserting a custom IP Header and informing the operating system that a header is not

needed. This process creates spoofed IP datagrams that may be used to impersonate another source IP

address or hide the true source IP address. These custom packets can created by using raw sockets in UNIX

-like systems or packet drivers such as WinPcap for Windows.

Blind/Nonblind Spoofing

Spoof attacks are generally considered blind attacks because the attacker will not recieve replies

directly from the victim. The victim usually only replies back to the source address of the recieved

packets and does not know the true source address of spoofed packets. If the attacker is on the same

subnet of the host being impersonated, then the attacker can "sniff" the reply and intercept the

packet. Scenarios where the attacker inspects the responses from the victim are called "nonblind spoofing"

and in other cases it is called "blind spoofing" where the responses are not intercepted.

Sequence Number Prediction

Once a connection between two hosts is established by using TCP, the exchange of packets between the two

parties also carry sequence numbers and acknowledgments. The sequence numbers are used to determine lost

packets or the order of packet and provide the reliability of TCP. Sequence numbers are generated pseudo-

randomly by an algorithm known by both parties. This prevents spoofing attacks from succeeding unless the

correct next sequence numbers are attached to the spoofed packets. If the responses can be viewed, the

attacker may be able to determine the algorithm and send packets with the correct sequence numbers. Once

this is accomplished the attack may intercept the session.

Denial of Service Using IP Spoofing

TCP connections are setup using three-way handshake that involve the SYN and ACK flags in

the TCP header. Connection that have been initiated but not completed are called half-open connections.

The state of half-open connections are stored and take up part of a finite-sized data structure. If a

spoofed SYN packet is sent to initiate a three-way handshake to a victim, the victim sends a

SYN-ACK packet and waits for the final ACK to complete the handshake. If the spoofed address

does not belong to the attacker, then the connection is left half-open and waiting indefinitely or for a

period of time. Having too many half-open connections will flood the finite-size data structure and

prevent and new connections from forming. This causes the victim to deny service to any legitimate

connections.

Session Hijacking

Using a combination of Sequence Number Prediction and Denial of Service attacks, an attacker can hijack a

session from a trusted host. Once the sequence number has been predicted, the attacker can send a reset

message to one of the two trusted hosts to end the connection from one host by impersonating the other.

The attacker can then impersonate the trusted host that dropped the connection by impersonating the source

address. At this point the attacker can preform commands with the priviledges of the dropped host. Figure 2 shows a visual representation of session hijacking.

Image:sessionhijacking.jpg

References

The Internet Protocol Journal, Volume 10, No. 4] --Luongp 22:59, 12 April 2009 (EDT)

Personal tools