ERASURE OF INTERNET BULLETIN BOARDS COULD BE HATE ATTACK.

Byline: Rory J. O'Connor Knight-Ridder Tribune News Wire

Vandals swept through the Internet last weekend, wiping clean dozens of public bulletin boards used by groups of Jews, Muslims, feminists and homosexuals, among others.

In one of the most widespread attacks on the international computer network, the programs automatically erased copies of more than 27,000 messages from thousands of servers, before operators stopped the damage.

The identity of those responsible for launching the apparent hate attacks is unknown.

The incident further illustrates the shaky security foundation of the Internet, which has mushroomed from academic research tool to international communications medium in just three years.

And it raised the ire of many Internet users furious at the ease with which a user can erase someone else's words from worldwide discussion groups, known as Usenet newsgroups, in a matter of hours.

``There's nothing you can do as an individual user to prevent someone from canceling your message,'' said John Gilmore, a computer security expert in San Francisco. ``We need something added to Usenet's software that would only allow a cancellation from the originator.''

The incident follows closely three other well-publicized Internet attacks.

In two cases, hackers altered the World Wide Web home pages of the Justice Department and the CIA, apparently as political protests. In the third, a hacker overloaded the computers of an Internet service provider called Panix with hordes of phony requests for a connection, thus denying use of the service to legitimate users.

The latest attacks - called cancelbots - were launched sometime over the weekend from a variety of Internet service providers, including UUNet Technologies in Fairfax, Va., and Netcom Inc. in San Jose. One attack was launched from a tiny provider in Tulsa, Okla., called Cottage Software, according to its owner, William Brunton.

``The offending user has been terminated and the information has been turned over to the proper (federal) authorities,'' Brunton said in a telephone interview Wednesday. ``It's now in their hands.''

Legal experts said it's unclear if the attacks constitute a crime under federal laws such as the Computer Fraud and Abuse Act.

``It's really a difficult issue,'' said David Sobel, legal counsel of the Electronic Privacy Information Center in Washington. ``Can you assign value to a newsgroup posting? Because most of the computer crime statutes assume you're ripping off something of value.''

While some of the deleted messages have been restored on certain servers, where operators have retrieved them from backup copies of their disks, users of other servers where the messages haven't been restored will never be able to read them.

The fact that a user can stamp out the words of someone else is an artifact of the original design of the Internet, begun as a Department of Defense project in 1969.

The Internet consists of tens of thousands of computers, called servers, that act as repositories for public messages, private electronic mail and World Wide Web home pages. Servers throughout the world are interconnected through telephone lines so they can exchange information and route messages to the individual users, or clients, of a given server.

Each server stores a copy of the constantly changing contents of newsgroups, which function as giant electronic bulletin boards dedicated to particular subjects. There are thousands of them, covering everything from particle physics to soap operas.

Any Internet user is free to post a contribution to nearly any newsgroup, and the posting is rapidly copied from one server to another, so the contents of a newsgroup are identical on every server.

Almost the only form of control over postings, including their content, is voluntary adherence to informal behavior rules known as ``netiquette.''

The idea of cancelbots originated when the Intern`et and its newsgroups were almost exclusively the domain of university and government scientists and researchers. Their purpose was to allow individuals to rescind messages they later discovered to contain an error. The action took the form of an automatic program, itself in the form of a message, because it would be impossible for an individual to find and delete every copy of the posting on every Internet server.

But the Usenet software running on servers doesn't verify that the cancel message actually comes from the person who created the original posting. All a malicious user need do is replace their actual e-mail address with that of someone else to fool Usenet into deleting a message. That counterfeiting is as simple as changing an option in the browser software most people use to connect to the Internet.