Security Information

DOS Attacks: Instigation and Mitigation

During the release of a new software product specialized to track spam, ACME SoftwareInc notice that there was not as much traffic as they hoped to receive. During furtherinvestigation, they found that they could not view their own website. At that moment, theVP of sales received a call from the company's broker stating that ACME Software Incstock fell 4 point due to lack of confidence. Several states away, spammers didn't like theidea of lower profit margins do to an easy to install spam blocking software so theythought they would fight back. Earlier that day, they took control of hundreds ofcompromised computers and used them as DoS zombies to attack ACME Software Inc'sInternet servers in a vicious act of cyber assault. During an emergency press conferencethe next morning, ACME Software Inc's CIO announced his resignation as a result of aseveral million dollar corporate loss.

Scenarios like the one above happen a more then people think and are more costlythen most will admit. Denial of Service (DoS) attacks are designed to deplete theresources of a target computer system in an attempt to take a node off line by crashing oroverloading it. Distributed Denial of Service (DDoS) is a DoS attack that is engaged bymany different locations. The most common DDoS attacks are instigated through virusesor zombie machines. There are many reasons that DoS attacks are executed, and most ofthem are out of malicious intent. DoS attacks are almost impossible to prevent if you aresingled out as a target. It's difficult to distinguish the difference between a legitimatepacket and one used for a DoS attack.

The purpose of this article is to give the reader with basic network knowledge abetter understanding of the challenges presented by Denial of Service attacks, how theywork, and ways to protect systems and networks from them.


Spoofing - Falsifying an Internet address (know as spoofing) is the method an attackeruses to fake an IP address. This is used to reroute traffic to a target network node or usedto deceive a server into identifying the attacker as a legitimate node. When most of usthink of this approach of hacking, we think of someone in another city essentiallybecoming you. The way TCP/IP is designed, the only way a criminal hacker or crackercan take over your Internet identity in this fashion is to blind spoof. This means that theimpostor knows exactly what responses to send to a port, but will not get thecorresponding response since the traffic is routed to the original system. If the spoofing isdesigned around a DoS attack, the internal address becomes the victim. Spoofing is usedin most of the well-known DoS attacks. Many attackers will start a DoS attack to drop anode from the network so they can take over the IP address of that device. IP Hijacking isthe main method used when attacking a secured network or attempting other attacks likethe Man in the Middle attack.

SYN Flood - Attackers send a series of SYN requests to a target (victim). The targetsends a SYN ACK in response and waits for an ACK to come back to complete thesession set up. Instead of responding with an ACK, the attacker responds with anotherSYN to open up a new connection. This causes the connection queues and memory bufferto fill up, thereby denying service to legitimate TCP users. At this time, the attacker canhijack the system's IP address if that is the end goal. Spoofing the "source" IP addresswhen sending a SYN flood will not only cover the offender's tracks, but is also a methodof attack in itself. SYN Floods are the most commonly used DoS in viruses and are easyto write. See

Smurf Attack- Smurf and Fraggle attacks are the easiest to prevent. A perpetrator sends alarge number of ICMP echo (ping) traffic at IP broadcast addresses, using a fake sourceaddress. The "source" or spoofed address will be flooded with simultaneous replies (SeeCERT Advisory: CA-1998-01). This can be prevented by simply blocking broadcasttraffic from remote network sources using access control lists.

Fraggle Attack - This types of attack is the same as a Smurf attack except using UDPinstead if TCP. By sending an UDP echo (ping) traffic to IP broadcast addresses, thesystems on the network will all respond to the spoofed address and affect the targetsystem. This is a simple rewrite of the Smurf code. This can be prevented by simplyblocking broadcast traffic from remote IP address.

Ping of Death - An attacker sends illegitimate ICMP (ping) packets larger than 65,536bytes to a system with the intention of crashing it. These attacks have been outdated sincethe days of NT4 and Win95.

Teardrop - Otherwise known as an IP fragmentation attack, this DoS attack targetssystems that are running Windows NT 4.0, Win95 , Linux up to 2.0.32. Like the Ping ofDeath, the Teardrop is no longer effective.

Application Attack - Thess are DoS attacks that involve exploiting an applicationvulnerability causing the target program to crash or restart the system.

Kazaa and Morpheus have a known flaw that will allow an attacker to consume allavailable bandwidth without being logged.See

Microsoft's IIS 5 SSL also has an easy way to exploit vulnerability. Most exploits likethese are easy to find on the Internet and can be copied and pasted as working code.There are thousands of exploits that can be used to DoS a target system/application. See

Viruses, Worms, and Antivirus - Yes, Antivirus. Too many cases where the antivirusconfiguration is wrong or the wrong edition is installed. This lack of foresight causes anunintentional DDoS attack on the network by taking up valuable CPU resources andbandwidth. Viruses and worms also cause DDoS attacks by the nature of how theyspread. Some purposefully attack an individual target after a system has been infected.The Blaster worm that exploits the DCOM RPC vulnerability (described in MicrosoftSecurity Bulletin MS03-026) using TCP port 135 is a great example of this. The Blastertargeted Microsoft's windows update site by initiating a SYN FLOOD. Because of this,Microsoft decided to no longer resolve the DNS for ''.

DoS attacks are impossible to stop. However, there are things you can do tomitigate potential damages they may cause to your environment. The main thing toremember is that you always need to keep up-to-date on the newest threats.


Antivirus software - Installing an antivirus software with the latest virus definitions willhelp prevent your system from becoming a DoS zombie. Now, more then ever, this is animportant feature that you must have. With lawsuits so prevalent, not having the properprotection can leave you open for downstream liability.

Software updates - Keep your software up to date at all times. This includes antivirus,email clients, and network servers. You also need to keep all network Operating Systemsinstalled with the latest security patches. Microsoft has done a great job with makingthese patches available for their Windows distributions. Linux has been said to be moresecure, but the patches are far more scarce. RedHat is planning on incorporating theNSA's SE Linux kernel into future releases. This will give Mandatory Access Control(MAC) capabilities to the Linux community.

Network protection - Using a combination of firewalls and Intrusion Detection Systems(IDS) can cut down on suspicious traffic and can make the difference between loggedannoyance and your job. Firewalls should be set to deny all traffic that is not specificallydesigned to pass through. Integrating an IDS will warn you when strange traffic is presenton your network. This will assist you in finding and stopping attacks.

Network device configuration - Configuring perimeter devices like routers can detectand in some cases prevent DoS attacks. Cisco routers can be configured to activelyprevent SYN attacks starting in Cisco IOS 11.3 and higher using the TCP interceptcommand in global configuration mode.

Access-list number {deny | permit} tcp any destination destination-wildcardip tcp intercept list access-list-numberip tcp intercept ? (will give you a good list of other options.)

Cisco routers can prevent Smurf and Fraggle attacks by blocking broadcast traffic. SinceCisco IOS 12.0, this is the default configuration. ACLs or access control lists should alsobe configured on all interfaces.

No ip directed-broadcast

The Cisco router can also be used to prevent IP spoofing.ip access-group list in interfaceaccess-list number deny icmp any any redirectaccess-list number deny ip anyaccess-list number deny ip anyaccess-list number deny ip host anySee Improving Security on Cisco Routers -

Old Cisco IOS versions are vulnerable to several DoS attacks. The "Black Angels" wrotea program called Cisco Global Exploiter. This is a great software to use when testing thesecurity of your Cisco router version and configuration and can be found at

Security is not as mystical as people believe. DoS attacks come in many differenttypes and can be devastating if you don't take the proper precautions. Keep up to date andtake steps to secure network nodes. Keeping security in mind can minimize damages,downtime, and save your career.

Security Resources:
Black Angels:
Forum of Incident Response and Security Teams:
SANS Institute:

Author: Jeremy Martin CISSP, ISSMP, ISSAP, CEI, CEH, CHS-III, CCNA, Network+, A+

Member of:
BECCA - Business Espionage Controls & Countermeasures Association
ISACA® - Information Systems Audit and Control Association
(ISC)² - International Information Systems Security
Certification ConsortiumISSA - Information Systems Security Association.
OISSG - Open Information Systems Security Group
YEN NTEA - Young Executives Network

could not open XML input