The public SMTP network is lousy with abusive spammers who would ruin the experience for the rest of us. Security Center Editor Larry Seltzer wonders whether taking some of your e-mail back onto private networks is the way to cut out the spam.
People with money and influence, and who dont care for the outside element, can fence off their community from the rest of the world. Outsiders can get in, but only by passing through a security gate. A phenomenon similar to this is developing in the world of e-mail, especially corporate e-mail, and it could help limit spam for those involved.
The best candidates for such an implementation are the large, hosted spam services such as Postini
The way these services work is that incoming e-mail passes through their servers and is filtered before it is sent to the clients servers. Technically, they hold the MX records for the clients mail domain.
At the same time, these companies have more than one client. Consider that since the spam services already have a private, trusted relationship routing e-mail to their client companies, any mail passing between them could be considered trusted. Theres no real need to pass this mail through filters, assuming these clients are responsible organizations with some sort of internal e-mail controls.
Its sort of like a big white list, but more sophisticated because it looks at the routing of mail. Suppose I white-list the address email@example.com because I buy my wife flowers now and then, and thats the online flower shops confirmation address (note to self: buy the wife some flowers). A spammer could get through to my Inbox by spoofing that address in the From: field. But if my gateway spam blocker would check the referring server against a white list of clean servers, my blocker wouldnt be fooled by a simple From: spoof attempt.
Now, a spam message with truly-spoofed headers would be another matter. Still it would be very hard to get an outside message through this network and have it appear as of it was completely internal. In fact, I doubt anyone could do it.
At the same time, my point about responsible organizations with internal controls is really important. Some of the hosted services work with ISPs and other public mail services, the most famous being Hotmails use of Brightmail
for spam filtering.
When you sign up for an ISP or a service like Hotmail you typically agree not to send spam, but there are individuals who would violate such sacred agreementsshocking as this may be to some of youwith no concern for the consequences, such as the loss of their free e-mail account. This is why private e-mail networks may prefer to work with the corporate networks rather than conventional ISPs. Incidentally, Brightmail supports scanning of outbound mail, although its not something every customer would implement.
This type of private network will really only be useful in direct proportion to the number of participants, so it would only begin to make sense if the spam filtering company has lots of customers. Theyd also all probably have to consent to being part of this private community.
The point of private networksat least intelligently designed onesis to define safe zones for e-mail, virtual spaces where you can relax your guard, and lower the risk of false positive identifications of messages. You still have to let mail in from the outside world because thats what makes e-mail useful. But at the same time, you can apply enhanced scrutiny to that mail. Private networks may not be a solution to spam, but for some folks it certainly could help.
Security Supersite Editor Larry Seltzer has worked in and written about the computer industry since 1983.
More from Larry Seltzer