Miscellaneous attacks against the Internet have become one of the major concerns
among many imperfectionsin contemporary cyberspace. In particular, worms are one highly
dangerous type. Howto timely detect and utterly stop such attacks still remains an
unresolved issue, and it is of greatnecessity to review the existing quality work in the
field of worm detection. In this paper, majorcharacteristics, working mechanisms, and
the life cycle of worms are introduced. Then, we divideexisting worm detection
techniques into two categories: machine learning methods and traditionaldetection
methods, and existing worm countermeasures are reviewed in detail by category.
Finally,some open issues in the field of worm detection are discussed, and the contents
of this paper aresummarized. The goal of this paper is to let researchers
comprehensively understand various existingdefense mechanisms against worms.