Wham, bam

Eradicating spam is never easy, and some may doubt it can ever be done. But that goal may be nearer than we think, thanks to a new spam technology dubbed PerfectMail…at least that’s what its makers claim.

The new entrant to the spam blocking marketplace isn’t what it seems, says Brampton, Ont.-based, XPM Software, who developed PerfectMail.

Their application, they say, is different from other anti-spam products in that it uses a peer-to-peer e-mail pattern recognition, dubbed Adaptive Learning to separate the wheat from the chaff.

“We tell people we are in the anti-spam market because that’s what [they] relate to,” said Larry Karnis, president of XPM Software. “We are not in the anti-spam market we are in the legitimate e-mail market.”

An appliance-based product, PerfectMail acts as an e-mail proxy firewall that sits at the edge of the network. To the outside world it’s the de facto mail server, said Karnis. It accepts all e-mail traffic, legitimate and malicious, virus laden or otherwise, separates the good from the bad, and sends only legitimate e-mail to an internal mail server.

The internal mail server never sees the malicious traffic, Karnis said. Adaptive Learning, said Karnis, examines mail traffic, “learns” who legitimate senders are, and then readjusts its behaviour so those senders’ e-mails are never blocked.

PerfectMail’s management interface lets administrators analyze how e-mail is being used, review e-mail traffic patterns for servers and domains, and drill down to the individual user level to monitor employee usage in detail, according to the company.

“The general thinking is if I block all your spam, what is left is legitimate e-mail,” said Karnis. “The problem is if I am not 100 per cent accurate, I [also block] some percentage of legitimate e-mail. The perception being, that is the price you pay for keeping the trash out of your in basket.”

Outside of the normal control flow of e-mail (read bad e-mails from good friends) the software offers black and white lists that allow previously acceptable senders to be added to a black list and e-mails from that marked person would no longer be received – but often times this can only be applied after the fact.

For example, if the user is going through a messy divorce she may not know to blacklist her custodial adversary after his lawyer updates him on her current demands.

“The problem is that the filter can’t make subjective decisions about who is nice and who is naughty,” said Karnis. “It only makes objective decisions. The filter needs to be told about changes to peer-to-peer e-mail relationships.”

Karnis says the black and white list is only a last resort because PerfectMail learns who the users’ peers are and can easily distinguish between spammers and legitimate mailers.

“You don’t normally need to work with a black and white list because the e-mail sender’s behaviour will put them in one category or the other,” says Karnis.

PerfectMail does not use quarantines. “I think quarantines are bad,” Karnis says. “If you send me an e-mail there is a somewhat greater than a zero per cent chance that we may incorrectly identify it. But we never quarantine it.”

Karnis says he likes to stick with the basics. “Are you being nice to me, am I being nice to you?” says Karnis. “If the answer is yes than the filters job is not to second guess but to facilitate.”

PerfectMail can affectively handle organizations with between 30 and 4,000 e-mail users, according to the company.

Related links:

Canadians winning the spam battle says poll

Spam? What spam?

Spam a nuisance, but does it really cost billions?