Skip to main content

2. Background

A simplified diagram of the TCP SYN flooding problem is depicted below:

                                                204.69.207.0/24
host <----- router <--- Internet <----- router <-- attacker

         TCP/SYN
     <---------------------------------------------
           Source: 192.168.0.4/32
SYN/ACK
no route
         TCP/SYN
     <---------------------------------------------
           Source: 10.0.0.13/32
SYN/ACK
no route
         TCP/SYN
     <---------------------------------------------
           Source: 172.16.0.2/32
SYN/ACK
no route

[etc.]

Assume:

  • The "host" is the targeted machine.
  • The attacker resides within the "valid" prefix, 204.69.207.0/24.
  • The attacker launches the attack using randomly changing source addresses; in this example, the source addresses are depicted as from within [4], which are not generally present in the global Internet routing tables, and therefore, unreachable. However, any unreachable prefix could be used to perpetrate this attack method.

Also worthy of mention is a case wherein the source address is forged to appear to have originated from within another legitimate network which appears in the global routing table(s). For example, an attacker using a valid network address could wreak havoc by making the attack appear to come from an organization which did not, in fact, originate the attack and was completely innocent. In such cases, the administrator of a system under attack may be inclined to filter all traffic coming from the apparent attack source. Adding such a filter would then result in a denial of service to legitimate, non-hostile end-systems. In this case, the administrator of the system under attack unwittingly becomes an accomplice of the attacker.

Further complicating matters, TCP SYN flood attacks will result in SYN-ACK packets being sent to one or many hosts which have no involvement in the attack, but which become secondary victims. This allows the attacker to abuse two or more systems at once.

Similar attacks have been attempted using UDP and ICMP flooding. The former attack (UDP flooding) uses forged packets to try and connect the chargen UDP service to the echo UDP service at another site. Systems administrators should NEVER allow UDP packets destined for system diagnostic ports from outside of their administrative domain to reach their systems. The latter attack (ICMP flooding), uses an insidious feature in IP subnet broadcast replication mechanics. This attack relies on a router serving a large multi-access broadcast network to frame an IP broadcast address (such as one destined for 10.255.255.255) into a Layer 2 broadcast frame (for ethernet, FF:FF:FF:FF:FF:FF). Ethernet NIC hardware (MAC-layer hardware, specifically) will only listen to a select number of addresses in normal operation. The one MAC address that all devices share in common in normal operation is the media broadcast, or FF:FF:FF:FF:FF:FF. In this case, a device will take the packet and send an interrupt for processing. Thus, a flood of these broadcast frames will consume all available resources on an end-system [9]. It is perhaps prudent that system administrators should consider ensuring that their border routers do not allow directed broadcast packets to be forwarded through their routers as a default.

When an TCP SYN attack is launched using unreachable source address, the target host attempts to reserve resources waiting for a response. The attacker repeatedly changes the bogus source address on each new packet sent, thus exhausting additional host resources.

Alternatively, if the attacker uses someone else's valid host address as the source address, the system under attack will send a large number of SYN/ACK packets to what it believes is the originator of the connection establishment sequence. In this fashion, the attacker does damage to two systems: the destination target system, as well as the system which is actually using the spoofed address in the global routing system.

The result of both attack methods is extremely degraded performance, or worse, a system crash.

In response to this threat, most operating system vendors have modified their software to allow the targeted servers to sustain attacks with very high connection attempt rates. This is a welcome and necessary part of the solution to the problem. Ingress filtering will take time to be implemented pervasively and be fully effective, but the extensions to the operating systems can be implemented quickly. This combination should prove effective against source address spoofing. See [1] for vendor and platform software upgrade information.

Last updated on