SpeedGate Spam FAQ

Spam Frequently Asked Questions

Here are some of the most commonly asked questions (and the answers) about SpeedGate mail accounts:

Information/How-To

The usual definition of spam is unsolicited commercial email, or unsolicited bulk email. For our purposes, that means two things:

It's trying to sell, advertise or promote something
You didn't request it.

Why is spam bad? Well, it's certainly annoying, but the key issue is that it shifts the burden of resources from the people sending it to the people receiving it. Unlike printed mail, spammers can send out thousands of messages for pennies -- messages which cost you time to download and to delete, which clog Internet traffic, and which use up our bandwidth and storage space.

It's also dishonest. Many spammers try to trick you into opening the message by using misleading subject lines, or force your email program to open a web browser. Many fake their return addresses so that complaints don't reach them -- or hit some other unsuspecting person instead. Because they can't use legitimate services, they try to piggyback on other people's mail servers -- literally stealing traffic from people and businesses with insecure servers. Some will even drive around looking for open wireless networks and use your internet connection to send their spam!

2004 saw the rise of two more obviously criminal trends: First, many spammers are now working with virus writers, using infected computers to do their dirty work, and launching denial-of-service attacks that have taken down several anti-spam services. And identity thieves are sending spam that pretends to be from banks, or eBay, or PayPal in order to trick people into handing over their credit card numbers, bank accounts, or even social security numbers.

Wading through 100 pieces of junk email to get at the three messages you need to run your business, or a "hello" from a friend, is patently ridiculous.

^ Contents

Do you allow customers to send spam? (No, we don't!)

SpeedGate's terms of service disallow the sending of unsolicited bulk email. You can send out newsletters to thousands of people if you want, but only to people who have signed up to receive them.

^ Contents

Do you sell your email address list? (No, we don't!)

Absolutely not! Why should we help people send mail to our customers that they don't want to receive?

^ Contents

Do you block incoming spam?

As much as we can!

We use blacklists to reject connections from sites known to send or relay spam. Then our server uses several pieces of software to (1) identify common spam-like traits, (2) compare against more blacklists for scoring instead of outright blocking, and (3) take a fingerprint of each message and compare it to databases of known spam. Mail exceeding a certain score is labeled in the subject and headers, and mail that exceeds a higher score is rejected. We also look for and attempt to block dictionary attacks.

The only blacklists we use to reject mail are the Spamhaus SBL+XBL and our own local list of sites that have recently sent us spam or viruses. A number of others contribute to the overall spam score.

Because no automated system can be perfect, we only label suspected spam instead of deleting it outright unless we can be reasonably certain that it is, indeed spam. At present those criteria include:

The message's fingerprint is found in the Razor or Pyzor databases and scores at least 10
The sending server is listed on at least three blacklists
The message advertises a website listed on at least three SURBL-type lists and scores at least 10.
The total score is above 15.

Anything else is labeled in the subject line and given an extra X-Spam-Score header with the score as a row of asterisks to make it easy for you to filter it in your email program:

X-Spam-Score: 13.6 (*************)

In the event that a message sent to you is incorrectly blocked, the sender will receive a notice explaining why -- whether it was blocked because of its source or because its structure and content looked too much like spam. It will never be silently discarded.

^ Contents

What does the spam score mean?

SpeedGate uses a program called SpamAssassin to scan mail for common spam traits. Each of several hundred tests will raise or lower the message's total score, and when it reaches 5 or higher, SpamAssassin will tag the message as spam by adding the label "[spam score x.x]" to the subject. Most will be allowed through with this label, but those that match a database of known spam will be rejected outright with a notice to the sender. You can have your email program filter the tagged messages into a separate folder or straight to the trash.

You can get a detailed report on just why SpamAssassin thought your mail was spam by looking at the full headers of the message (unless you are using Eudora, in which case the spam report is always displayed -- but you can change this if you want to). For directions, see How do I see the "full headers" of a message?

Note that no mail will ever be dropped without the sender being notified (assuming that the sender uses a valid email address).

The exact rules used to identify spamlike traits are too numerous to list here, but will be made available on request.

^ Contents

What does all this stuff in X-Spam-Status mean?

On many of your messages you may see something like this in the headers:

X-Spam-Status: No, hits=-12.2 required=5
	tests=BAYES_10,CLICK_BELOW,GENUINE_EBAY_RCVD,HTML_50_60,
	      HTML_FONT_BIG,HTML_FONT_COLOR_GRAY,RCVD_IN_BONDEDSENDER
	autolearn=ham version=2.53

This is extra diagnostic information, mainly useful for figuring out why a spam message slipped through the cracks, so you shouldn't have to worry about it. However, if you're curious:

Yes/No: Self-explanatory -- whether SpamAssassin thought this was spam or not.
hits: The final spam score for the message.
required: The minimum score required for a message to be labeled as spam. This is constant right now, but we may add the ability to change this on a per-user basis sometime in the future.
tests: A condensed listing of all the rules that the message matched. Some of these increase the score, and others decrease it. If the message was labeled as spam, you should have a full report also attached. If not, you can look up descriptions and default scores for each test at SpamAssassin/tests.
autolearn: SpamAssassin uses a method called a Bayesian classifier which actually learns from the kinds of spam and non-spam that passes through our server. If the score is very high or very low, it will study it as "spam" or "ham" (a pun on the original meaning of "SPAM"). This will only appear if SpamAssassin auto-learned from this message.
version: The version of SpamAssassin that ran the checks.

Problems

I just signed up and I'm already getting spam! I thought you didn't sell your mailing list!

As we said before, we don't. Chances are one of two things has happened to you: either a previous customer had your address, or you have fallen prey to a dictionary attack.

We all know how hard it is to get off a spam list. Unfortunately, even addresses that don't exist anymore tend to be stuck there for a long time. (Ever notice how many dead links are on the web? Same principle.) So let's say for instance that you have just signed up as troll@speed.net. Now suppose that two years ago, we had someone else who was troll@speed.net, who did everything -- posted to newsgroups, put his address on his website, signed up for newsletters from other people who did sell his address -- and ended up on dozens of spammers' lists before cancelling a year ago. Some of those lists will have dropped the address off due to it bouncing for a year. But some of them -- whether published on CD-ROM or just poorly maintained -- kept it. So when we create a new troll@speed.net, the account just stops bouncing.

Then there's the dictionary attack, also known as the Rumplestiltskin attack (because the method is guessing names). Spammers will take a list of common names, combine them with first or last initials, and just send a message to all such addresses on a particular server. Sometimes they'll watch to see what bounces and only keep the ones that don't. But there's always someone else who hasn't done that yet.

^ Contents

Hey, I just got spam from speed.net!

First, check the source carefully. Does it actually say it came from/through speed.net, or does it say it came from someplace like speed.net.id? We have gotten several complaints of spam that actually came from speed.net.id, which appears to be some company in Indonesia (we're based in the United States).

Second, we've been getting reports of (and bounces from) spam which has forged a fake speed.net return address. Please see our forged spam page for more details.

Finally, if the spam definitely came from the speed.net network, please tell us! Send a message to abuse@speed.net and we'll track down the person responsible!

^ Contents

My favorite newsletter got labeled as spam!

Unfortunately, many newsletters share characteristics with spam. It all boils down to whether you signed up for them or not. If you get mail that's labeled as spam but isn't, let us know at spam-false-positive@speed.net, and we can usually alter the filter to skip those messages. It's especially easy with newsletters that always come from the same address.

^ Contents

I tried to send email and you said it was spam!

We try very hard not to actually block legitimate email. As described in our section above on blocking spam, we only reject apparent spam under strict circumstances. Most of the time, if a legitimate message is identified as spam, it will only be labeled and will still reach its destination.

If the message you sent was blocked, you should look at the rejection message to see why:

"No mail accepted from known spam hosts -- see <URL>" indicates that the server you sent through has been identified as a source of spam. Please send the information including the URL to whoever manages your mail server so that they can determine why the server was listed and go about getting your site removed.

"No mail accepted from known insecure relays -- see <URL>" indicates that the server you sent through has been determined to be running an open relay and is exploitable by spammers. This is kind of like putting your phone outside the house and letting anyone use it for free without asking. Contact the person/department/company that manages your mail server and provide them with this information and the URL listed so that they can secure the server and get removed from the list.

In both cases, if you have difficulty getting off of the list, contact us again (use the postmaster address or our contact form) and we can set up an exception for your network.

"Message seems to be spam" with no URL indicates that our content analysis found an extremely large number of spam-like traits in your message, or that it found a medium number of traits and found the message's fingerprint in a database like Razor or Pyzor. To give you an idea, 5 points gets a message labeled as spam, but we don't reject until it hits 25. See Do you block incoming spam? above for a more detailed explanation.

^ Contents

Send comments on our web pages to: webmaster@speed.net

Site Features

Section

Quick Jump

Spam Frequently Asked Questions

Information/How-To

Problems