"Neither snow nor rain nor heat nor gloom of night stays these couriers from the swift completion of their appointed rounds."
--Inscription on New York City Post Office
The quotation above is a loose translation taken from The Histories of Herodotus, describing the famous Persian couriers who used a relay system of horses and riders to deliver messages across the Persian Empire. (William M. Kendall, an employee of the architectural company that designed the New York City Post Office building in 1914, suggested that this inscription be placed on the new building as a dedication to the postal workers who would inhabit it.)
In the world of electronic messages, we expect a similar resiliency from our SMTP mail servers: We expect our email to arrive swiftly through the Internet, regardless of the obstacles that might attempt to interfere with its delivery along the network. And, indeed, the very nature of TCP/IP, with its packet addressing schemes and legions of routers and servers, allows most of our email to arrive flawlessly anywhere in the world, almost instantaneously. Packets of email messages get relayed thousands and hundreds of thousands of times, enabling us to conduct our business communications in a more efficient and virtual manner.
But now all of that is in peril. The great flood of unsolicited email, called spam, has forced network administrators to take drastic draconian measures, and the result of those measures may quite soon destroy the Internet's email function. In fact, those measures may be impacting your email system right now.
DNS Blocking Targets Spammers but Also Hits Legitimate Users
Internet network administrators have begun using a technique called DNS blocking to cut off email that they believe is coming from spammers. They have developed blacklists of IP addresses that are known to be the sources of spam. These blocking blacklists are being promulgated by administrators throughout the Internet, so when an email passes through the network, the IP address of the sender's server is compared to addresses of servers believed to have been the source of spam in the past. If the addresses match, the email is rejected--sometimes sending a rejection message back to the originator and sometimes not. The purpose of these DNS blocks is to reduce the volume of spam, much as a traffic cop might pull over an unsafe delivery vehicle. They are becoming "filters" embedded within the Internet itself, fully automated and relatively bulletproof.
"So, no problem," you're thinking. "We don't send spam, so why should we worry?"
You should be concerned because the methods by which these automated administration mechanisms are identifying spamming servers may not necessarily be accurate or appropriate. The implementation of these services will sometimes interrupt the delivery of legitimate messages--sent by your SMTP mail server. And once your mail server becomes blocked, it becomes a nightmare to restore its former functionality.
Not only that, as more and more IP addresses of SMTP servers are added to the blacklist, our ability to send email may soon no longer be as resilient as we need in order to transact business on the Internet. In fact, sending a message via email may end up as error-prone and haphazard as communicating over twisted twine between two tin cans.
How can this happen? To get the answer, you need to understand a bit about SMTP and the functionality of something called open relays.
SMTP Open Relays
Simple Mail Transfer Protocol (SMTP) is the protocol that forms the backbone of our email communications. It's a protocol that consists of headers and tags identifying the elements of each email transaction such as From, To, Subject, and the Body of the message as well as the form and placement of text, file attachments, and other things. Email servers transmit and receive messages using the SMTP format, and email client software interacts with an SMTP server when it needs to send a message out along the Internet. (This is one protocol used in Internet email; the other is called Post Office Protocol (POP3), which is used to extract messages from an email account and offload those messages into the client software. Thus, as email users, we "send" messages using SMTP, but we generally receive them on our PCs through POP3 servers.)
One of the features of SMTP servers has always been--like the Persian couriers in the time of Herodotus--to relay messages passed to it by third parties to other SMTP servers on their routes. When an SMTP server is functioning in this capacity, it is called an open relay. In an open relay environment, it doesn't matter where the message originates: The purpose of the relay is to get the message to its destination. In such an environment, "Neither snow nor rain nor heat nor gloom of night stays these couriers from the swift completion of their appointed rounds." Just like the fabled Persian couriers of old or the not-so-fabled United States postal workers.
Unfortunately, spammers have learned that they can hijack the SMTP servers on the network to also indiscriminately relay their unsolicited email to anyone they want. As a result, if your SMTP server is configured as an open relay, it may be an unwitting accomplice to the flood of spam that is currently filling email mailboxes. How?
A spammer first tracks down an email address back to the originating IP address of the SMTP server. He then attempts to send a message back to himself through that SMTP server. If he is successful, he will have discovered an "open relay" and is free to utilize that server to send spam anywhere he chooses.
Meanwhile, if a recipient of his unsolicited email complains about the spam, the spammer is fully protected because his originating IP address will have been doctored to reflect the SMTP server that performed the open relay.
Following the Bouncing Email
Now, if an Internet network administrator puts the offending SMTP server on a DNS blocking blacklist--to shut off the source of the spam--the spammer simply looks for yet another open relay on some other SMTP server and continues on his merry way.
However, for legitimate users who need that SMTP server to deliver mail, the DNS block will also suddenly reject that legitimate outgoing email. The innocent sender of a legitimate email will have no recourse except to complain to his email administrator, who typically has few resources to fix the problem. This administrator will look at the rejected email and perhaps see a cryptic message like this:
Diagnostic-Code: smtp;550 5.7.1 Mail from 220.127.116.11 refused by blackhole site relays.osirusoft.com.
In the message above, the IP address 18.104.22.168 that represents the SMTP server has been blocked from delivering any mail along the route serviced by the blacklist service called "relays.osirusoft.com." To unblock the SMTP server, the server administrator must now contact this organization to remove the server from the blacklist.
However, this is merely one blacklist service that may be blocking email from the SMTP server. And the blacklist itself may have been already passed to other DNS blocking services, magnifying the problem. And, by the time an email administrator is made aware of the problem (when his users start complaining), the blacklist may have been passed to hundreds of DNS blocking services along the Internet, slowly choking off the functionality of the SMTP server until at last nothing can be delivered.
Automated DNS Blocking Services Are Closing Down All Open Relays
Most insidious is the fact that some DNS blacklisting services have begun automated attacks against all open relay SMTP servers, hunting them down in exactly the same manner that spammers have tracked them and then automatically adding those IP addresses to their blacklists. In other words, these DNS blacklisting services are actively shutting down every open relay in an effort to curtail the volume of spam, regardless of whether the particular SMTP server was ever used as a spamming source. Furthermore, some of these services are blacklisting entire ranges of IP addresses--not just a unique SMTP server address--in their zeal to choke off the spam.
Is this a legal use of the Internet? Who knows? But it's happening! And if your SMTP server hasn't been blocked, you need to protect it now. How?
The safest thing to do is to make certain that only registered users with email accounts at known IP addresses can use your SMTP server. This can usually be accomplished in the configuration of the SMTP server itself, closing down the server's open relay ability.
Legit Open Relays
Unfortunately, shutting down the SMTP server's ability to function as an open relay may also severely limit the use of SMTP by those who have a legitimate need to relay email. Consider, for instance, the salesman who is traveling with a laptop: He may have a need to relay email from the road using a dialup ISP. (For these users, implementing a "POP3 before SMTP" script may need to be implemented. This technique allows the SMTP server to send email for a brief time after the user has logged onto his POP3 email account and extracted messages.)
Unblocking the DNS Block
First, the email administrator needs to find out which DNS blocking services are actually blocking the SMTP server. This may be one service or dozens or hundreds. Fortunately, there are several sites on the Internet that will tell him where the server has been blacklisted, including http://relays.osirusoft.com and http://www.abuse.net. These services may allow you to run a query to determine where a DNS block has been put into effect. However, they are not particularly "user friendly," and they are replete with disclaimers about what they will or can do to help.
Next, the email administrator will be required to button down your SMTP mail server so that it no longer functions as an open relay. This means changing the configuration of the server so that only authorized users with email accounts can make use of it.
Then, after the open relay has been closed on your SMTP server, the email administrator will need to contact every one of the blacklisting services to have them retest your SMTP server for an open relay. Each blacklisting service has a different method of accomplishing this task, and it may take hours, days, or weeks for them to clear your site. Furthermore, once the block has been removed, it may take up to 24 hours for your service to be restored. In the meantime, the email users will no doubt complain that their email isn't getting through. They will be able to receive email through their POP3 accounts, but they'll be unable to respond back.
Changing Your IP Address May Not Work
Another technique to more quickly resolve a blocked SMPT server is to actually change the IP address of the SMTP server itself by contacting your hosting ISP. But if the server has dozens or hundreds of email clients tied to it, this too can take a tremendous amount of time, because each email client will need to be reconfigured to access the new IP address for mail.
Furthermore, if the SMTP server has been mistakenly blocked, it's possible that all the IP addresses associated with your Internet domain may have also been blocked, so the email administrator will need to check once again with the DNS blocking service to make certain the SMTP server has a clean bill of health.
Will DNS Blocking Stop Spammers?
Meanwhile, where is all of this DNS blocking leaving the functionality of the Internet? Will DNS blocking actually eliminate spammers? Probably not! Spammers will no doubt quickly create new schemes to find other security holes by which they can send their unsolicited bulk mail. It's only a matter of time.
But what is most troubling is the manner by which Internet administrators have dealt with the problem of spam: It's a preemptive and "blanket" condemnation, penalizing the innocent SMTP email installations as well as guilty spam sites. It kills all the open relays, without really addressing the root causes of Internet security. And though today these administrators are building automatic robots that are targeting spam sites, tomorrow they may build robots to target email clients that unwittingly proliferate viruses or, by extension, anything that they deem inappropriate. This technique, by any measurement of propriety, is an incursion into our businesses' infrastructures--without our companies' knowledge or consent. No warning is required: They can come and "fix" us as they please.
The result may be that we lose the flexibility and the functionality of the Internet as a business tool for sending communications. In our electronic analogy of Herodotus' Persian couriers, these Internet administrators may be shooting the messenger simply because they don't like the message that is being carried.