TLS Is a No-Brainer For Mail Servers

Opinion: There's no reason not to implement encryption between servers, even if it seems you're addressing a problem that doesn't exist.

Encryption has been an important feature of many Internet protocols for some time now. Transfers between mail servers havent historically been protected, but there is a protocol available to encrypt them, and everyone should now look to implement it.

TLS (Transport Layer Security) is easy to use. Of course, how easy it is depends on the specifics of each server, but generally its little more than clicking a checkbox in the servers administrative console. You do need to get a certificate, but you dont necessarily need to have one held by a certificate authority. You can generate your own certificate for free, and in many ways this works just as well.

The end result is that mail transfers on the Internet between servers will be encrypted, as opposed to the default of clear text. The protocol is set up to request an encrypted transfer and fall back to unencrypted if the recipient doesnt support encryption.

We know its easy to do, and as far as I can tell theres no downside. This is why secure managed e-mail provider Postini has announced that it will support TLS (which they call "Postini Auto-Encryption") for all outside mail interaction with their e-mail boundary service Perimeter Manager Enterprise Edition 5.2. Postini claims to be the fourth-largest e-mail provider, so this move should increase the amount of TLS in use on the Internet considerably.

/zimages/7/28571.gifClick here to read eWEEK Labs review of Postinis Perimter Manager.

So theres no reason not to implement TLS, but whats the positive rationale for it? What problem does it solve? The obvious answer—the one you find in press releases—is that it secures the connection between two servers through encryption. The implication is that, absent the encryption, someone could "tap" the connection between those servers and read the data.

Does this ever happen? I tried to find examples of it with no luck, but perhaps I didnt search cleverly enough. Perhaps, as one vendor suggested to me, it would be too embarrassing for a victim to admit, and so it remains secret. Or perhaps this isnt a real world problem.

Yet. I look at this solution to a non-problem and I see an opportunity to preempt a problem that could arise. The way problems creep up in this business its only too easy to imagine some protocol-level weakness or some mail server vulnerability cropping up and suddenly everybodys got to patch or implement TLS in a hurry.

Why not lock this door before someone walks in? TLS may be a solution in search of a problem, but on the Internet thats not always a bad thing.

Security Center Editor Larry Seltzer has worked in and written about the computer industry since 1983.

More from Larry Seltzer

/zimages/7/28571.gifCheck out eWEEK.coms for the latest security news, reviews and analysis. And for insights on security coverage around the Web, take a look at Security Center Editor Larry Seltzers Weblog.