Sequence Numbers - crucial to hijacking a session

Sequence Numbers

• 
  
  Sequence Numbers are very important to provide reliable communication but they are also crucial to hijacking a session.
• 
  
  Sequence numbers are a 32-bit counter, which means the value can be any of over 4 billion possible combinations.
• 
  
  The sequence numbers are used to tell the receiving machine what order the packets should go in when they are received.
• 
  
  Therefore an attacker must successfully guess the sequence number to hijack a session.

TCP provides a full duplex reliable stream connection between two end points. A connection is uniquely defined by the IP address of sender, TCP port number of the sender, IP address of the receiver and TCP port number of the receiver.

Every byte that is sent by a host is marked with a sequence number and is acknowledged by the receiver using this sequence number. The sequence number for the first byte sent is computed during the connection opening. It changes for any new connection based on rules designed to avoid reuse of the same sequence number for two different sessions of a TCP connection.

We have sent the increment of sequence number in our discussion of the three way handshake. What happens if the sequence number is predictable? When the TCP sequence is predictable, an attacker can send packets that are forged to appear to come from a trusted computer.

The next step taken was to tighten the OS implementation of TCP and introduce randomness in the ISN. This was done by the use of pseudo-random number generators (PRNGs). PRNGs introduced some randomness when producing ISNs used in TCP connections. However, adding a series of numbers together provided insufficient variance in the range of likely ISN values; thereby allowing an attacker to disrupt or hijack existing TCP connections or spoof future connections against vulnerable TCP/IP stack implementations.

This implied that systems relying on random increments to make ISN numbers harder to guess were still vulnerable to statistical attack. In other words, with the passage of time, even computers choosing random numbers will repeat themselves, because the randomness is based on an internal algorithm that is used by a particular operating system. Once a sequence number has been agreed to, all following data will be the ISN+1. This makes injecting data into the communication stream possible.

Threat

If a sequence number within the receive window is known, an attacker can inject data into the session stream or choose to terminate the connection. If the attacker knows the initial sequence number, he can send a simple packet to inject data or kill the session if he is aware of the number of bytes transmitted in the session this far.

As this is a difficult proposition, the attacker can guess a suitable range of sequence numbers and send out a number of packets into the network with different sequence numbers - but falling within the range. Since the range is known, it is likely that at least one packet will be accepted by the server. This way, the attacker need not send a packet for every sequence number, but resort to sending an appropriate number of packets with sequence numbers a window-size apart. But how does he know how many packets are to be sent?

This is obtained by dividing the range of sequence numbers to be covered by the fraction of the window size that is used as an increment. Why was this possible despite the introduction of PRNGs? The problem lay in the use of increments themselves, random or otherwise, to advance an ISN counter, making statistical guessing practical. The result of this is that remote attackers can perform session hijacking or disruption by injecting a flood of packets with a range of ISN values, one of which may match the expected ISN. The more random the ISNs are, the more difficult it is to carry out these attacks.