Common Types of Attacks
There are several things an intruder could do when he gets to your site. These are the most common.
Web defacement can take several forms, some subtle, others obvious. Most companies today have an Internet presence. Either it's hosted by an ISP or you're doing it yourself.
In building your Web site you have the copywriters rewrite your introduction and business model a dozen times just to get the language right. You've gone through at least as many revisions of your company logo and Web site art, not to mention the weeks you've spent tweaking the numbers so you can provide competitive products and services online.
Subtle defacement could take the form of altering your prices to make them 5% higher than your competitors or mixing up the product descriptions and prices.
Obvious defacement could be replacing your entire Web site with an image the intruder has hacked together of your CEO clubbing a baby seal. Another obvious defacement could be putting a message on your site describing how non-existent the security on your site was, and how no one ought to trust you with a customer's credit card information. Obviously, either of these could have a serious and significant impact on your business.
Unsolicited Commercial Email (UCE or Spam)
You don't have to be told that spam annoys a lot of people. It annoys the people who get it, it annoys the system administrators who spend a lot of time trying to block it, and it annoys the ISPs who are paying to store all these email messages on their servers.
If your mail server is configured to allow relaying of spam (by permitting unauthenticated mail to be relayed), not only does it place a burden on your mail server having to deliver all that additional email, but it could blacklist your domain. (We'll talk about how to guard against this later.)
Blacklisting is exactly what it sounds like. Most modern mail servers use several anti-spam controls, one of which is the Real-time Blackhole List or RBL. The Real-time Blackhole List prevents mail coming from known spamming domains from reaching your users. It does this by comparing the IP address of the sender with a list of known spammers. If the IP address is found, the mail is rejected.
The Real-time Black List is maintained by MAPS, the Mail Abuse Prevention System, a non-profit corporation at http://maps.vix.com/.
Allowing spam through your mail server could end up preventing your user base from sending valid email to anyone implementing spam control on their mail servers.
Spoofing an IP address is the act of replacing the real source address with a different source address. Because when the intruder uses a different or invalid IP address as the source, he won't see the reply, it's natural to wonder why he would bother. However, many attacks, such as Denial of Service (DoS) attacks, do not need to see the reply, they just need to get traffic to the intended target.
There are many reasons an intruder would spoof her IP address. First, you will not see the real IP address that the attack is coming from. Additionally, if the intruder uses an IP address internal to your network as the source address, your existing safeguards might allow that packet to pass.
If an intruder discovers an open security hole in your systems, he can then exploit that vulnerability to break into your system. This is quite different from an intruder breaking into your house. Unlike the home intruder, who must rush through your house before the movie is over, the intruder breaking into your system will usually take more time. Typically an intruder takes days or weeks to break into your system rather than hours. The extended break-in is used to obfuscate the attack in the background of otherwise normal activity.
The intruder gains access to your system usually by installing a series of Trojan-horse programs collectively known as a root kit. In the UNIX vernacular, root is the account with full, unbridled system privileges. Installing the root kit accomplishes two goals:
The Trojan programs allow normally untraceable access, so there is not as much sanitizing that the intruder must do to cover his tracks.
The intruder does not want to rely on a security vulnerability that you might eventually discover and close. The root kit allows him to create a more "legitimate" form of entry.
After the intruder has compromised one of your systems, he will often expand the attack by attempting to exploit any trust that system has with others and invade those as well. Frequently the intruder has no interest in your system other than to launch an attack onto another system and hide its origin.
Denial of Service (DoS)
A denial of service attack is a simple, but often extremely effective, attack that is difficult, if not impossible, to prevent. The goal of a denial of service attack is to deny access to your particular services, effectively preventing your organization from operating. A denial of service could be launched against any part of your Internet connectivity and network infrastructure.
An attacker could target your Internet connection, for example, with a DoS attack. By sending a simple flood ping, which barrages the target with ICMP Echo Request packets as fast as possible without waiting for replies, an attacker can cause more traffic than your Internet connection is capable of handling. This effectively prevents you from receiving or replying to legitimate requests. (ICMP is the Internet Control Message Protocol which is used to report errors to devices, usually routers. An echo request packet asks for an ICMP reply; sent continuously, your machine will get bogged down sending ICMP replies.)
Another type of Denial of Service attack can be caused by mail bombing, in which an attacker targets your email system by sending overly large email messages to users such as postmaster or Webmaster. This not only affects your Internet connection bandwidth, but also your mail server's capability to handle other email. An attack such as this could easily fill up your mail server's mail store drive, preventing you from sending or receiving new email.
TCP SYN Scan
Another common type of attack is the TCP SYN attack. Typically, when a TCP connection is initiated, the sending machine sends a SYN request, which is acknowledged by the recipient machine with an ACK, and the sending machine responds itself with an ACK. This three-way handshake sets the stage for a normal TCP connection.
Figure 3.1 shows a typical TCP connection.
The recipient computer uses memory resources waiting for the third part of the handshake, the ACK from the sender. A TCP SYN attack exploits this by providing a "spoofed" unreachable IP address in the SYN. Therefore the recipient computer responds with its ACK to a non- existent machine that never responds to the ACK:
Figure 3.2 illustrates a TCP SYN attack.
A TCP SYN scan performs large numbers of SYN connections in rapid succession against your servers, never sending the ACK response. This leaves hundreds, if not thousands, of connections half-open.
Distributed Denial of Service Attacks (DDoS)
Recent forms of distributed, coordinated, mass denial of service attacks have focused on the larger Internet targets such as financial companies, search engines, and government agencies. An intruder breaks into a large number of systems and leaves a root kit and client installed. Prepackaged denial of service "hacker tools" such as Trinoo/Tribal Flood, Stacheldraht, and Mstream are controlled remotely, allowing the attacker to later contact and coordinate attacks at will. When the attack comes, it will be from multiple innocent machines at once.
These distributed denial of service (DDoS) attacks cannot be prevented by the company being attacked. Individual companies, universities, and ISPs connected to the Internet are responsible for securing their networks. It is your duty to ensure your systems are not being used to attack others.