Home > Articles > Security > Network Security

Attacks from Outside the Operating System

  • Print
  • + Share This
This excerpt from Andy Tanenbaum's Modern Operating Systems looks at some of the operating systems aspects of external threats, primarily focusing on viruses, worms, mobile code, and Java applets.
From the author of

For machines connected to the Internet or another network, there is a growing external threat of attack. A networked computer can be attacked from a distant computer over the network. In nearly all cases, such an attack consists of some code being transmitted over the network to the target machine and executed there doing damage. As more and more computers join the Internet, the potential for damage keeps growing. In the following sections we will look at some of the operating systems aspects of these external threats, primarily focusing on viruses, worms, mobile code, and Java applets.

It is hard to open a newspaper these days without reading about another computer virus or worm attacking the world's computers. They are clearly a major security problem for individuals and companies alike. In the following sections we will examine how they work and what can be done about them.

I was somewhat hesitant to write this section in so much detail, lest it give some people bad ideas, but existing books give far more detail and even include real code (e.g., Ludwig, 1998). Also the Internet is full of information about viruses so the genie is already out of the bottle. In addition, it is hard for people to defend themselves against viruses if they do not know how they work. Finally, there are a lot of misconceptions about viruses floating around that need correction.

Unlike, say, game programmers, successful virus writers tend not to seek publicity after their products have made their debut. Based on the scanty evidence there is, it appears that most are high school or college students or recent graduates who wrote the virus as a technical challenge, not realizing (or caring) that a virus attack can cost the collective victims as much as a hurricane or earthquake. Let us call our antihero Virgil the virus writer. If Virgil is typical, his goals are to produce a virus that spreads quickly, is difficult to detect, and is hard to get rid of once detected.

What is a virus, anyway? To make a long story short, a virus is a program that can reproduce itself by attaching its code to another program, analogous to how biological viruses reproduce. In addition, the virus can also do other things in addition to reproducing itself. Worms are like viruses but are self replicating. That difference will not concern us here, so we will use the term ''virus'' to cover both for the moment. We will look at worms in Sec. 9.5.5.

9.5.1 Virus Damage Scenarios

Since a virus is just a program, it can do anything a program can do. For example, it can type a message, display an image on the screen, play music, or something else harmless. Unfortunately, it can also erase, modify, destroy, or steal files (by emailing them somewhere). Blackmail is also a possibility. Imagine a virus that encrypted all the files on the victim's hard disk, then displayed the following message:

GREETINGS FROM GENERAL ENCRYPTION!

TO PURCHASE A DECRYPTION KEY FOR YOUR HARD DISK, PLEASE SEND $100 IN SMALL, UNMARKED BILLS TO BOX 2154, PANAMA CITY, PANAMA. THANK YOU. WE APPRECIATE YOUR BUSINESS.

Another thing a virus can do is render the computer unusable as long as the virus is running. This is called a denial of service attack. The usual approach is consume resources wildly, such as the CPU, or filling up the disk with junk. Here is a one-line program that used to wipe out any UNIX system:

     main( ) {while (1) fork( );} 

This program creates processes until the process table is full, preventing any other processes from starting. Now imagine a virus that infected every program in the system with this code. To guard against this problem, many modern UNIX systems limit the number of children a process may have at once.

Even worse, a virus can permanently damage the computer's hardware. Many modern computers hold the BIOS in flash ROM, which can be rewritten under program control (to allow the manufacturer to distribute bug fixes electronically). A virus can write random junk in the flash ROM so that the computer will no longer boot. If the flash ROM chip is in a socket, fixing the problem requires opening up the computer and replacing the chip. If the flash ROM chip is soldered to the parentboard, probably the whole board has to be thrown out and a new one purchased. Definitely not a fun experience.

A virus can also be released with a specific target. A company could release a virus that checked if it was running at a competitor's factory and with no system administrator currently logged in. If the coast was clear, it would interfere with the production process, reducing product quality, thus causing trouble for the competitor. In all other cases it would do nothing, making it hard to detect.

Another example of a targeted virus is one that could be written by an ambitious corporate vice president and released onto the local LAN. The virus would check if it was running on the president's machine, and if so, go find a spreadsheet and swap two random cells. Sooner or later the president would make a bad decision based on the spreadsheet output and perhaps get fired as a result, opening up a position for you-know-who.

  • + Share This
  • 🔖 Save To Your Account