Computer worms propagate on their own. They search for other computers, probing the networks they are on either randomly or with some plan or design, such as computer names or addresses they find on a computer they have infected. When they find a computer, they attempt to make a network connection to it and take over; trying to find one or more vulnerabilities they can use to do so.
Worms really are often masterful, elegant, and expert examples of programming. They tie together the automation of services, daemons, or memory-resident viruses often with a level of strategy that almost suggests artificial intelligence. AI isn't beyond possible for worms of the future, either. They also require the programmer's understanding and use of vulnerabilities and exploits. The worms need a weakness in target computers out on the network in order to spread.
Part of what makes the Internet such a fertile ground for the spread of worms is the sheer volume of computers present, and the vast numbers of uneducated, inexperienced, or unconcerned users and computer owners or operators. Unfortunately, even one or a few vulnerable computers permitting a worm infection can paralyze a network. The worm can flood network segments, WAN circuits, and Internet connections with attempts to seek new victims and spread. For some historic worms this was the majority of the damage done--the worms didn't damage local computer programs or data, they simply clogged the network pipes wherever they went.
The main means of catching a worm is failure to have your anti-virus software signatures up-to-date. Most often anti-virus software that is kept up to date can stop a worm. Other weaknesses allowing you to catch a worm involve having no network firewall to block access by the worm to your computer network, and having no host-based firewall blocking access to each PC. Any one of these weaknesses could lead to worm infection. It's best practice to keep all of these means of protection in place and up to date. Let's look at protection in detail.