The answer to Verizon's mail blocking?

As an email postmaster in California, I found all our system's email to Verizon (and GTE) was being rejected. I found an answer on a user forum for the email system I have (CommuniGate - www.stalker.com). The reason turned out to be a new setting on our mail server, 'Delay SMTP Prompt'.

If turned on, and set for 30 seconds, it rejects any incoming SMTP connection that attempts to send data prior to the prompt being sent. The intent is to catch email viruses and spammers which blindly send email out. It works quite well... But you can never change just one thing. There are always unintended consequences.

When you send email to Verizon, their mail server tries to verify the sender's address. This connection is ALSO SMTP, and my server dutifully sent nothing until 30 seconds are up.

Unfortunately, Verizon seems to only wait 28 seconds for a response. If it doesn't get it within that timeframe, the email is rejected. Changing the delay to 28 seconds or less fixed our Verizon 'problem'. This may be the cause of what you have noted.

As to whether there is blocking of 'foreign IP addresses' on purpose, I can say there are a number of ways to do this using information at: www.iana.org/assignments/ipv4-address-space

For instance, depending on your mail server software, you could set up absolute rejection of email coming from or through alleged APNIC or RIPE IP address space. This does have obvious drawbacks. But then using Realtime Blocking Lists (RBLs) also has pros and cons.

If available on your system, you could include this characteristic in computing a total 'score' and 'filter' messages only if they exceed whatever you consider to be allowable. This may be better, but is still not perfect.

And regardless of what is done or not, someone will always complain. When doing any filtering there will always be false positives and false negatives, and complaints. But IMHO there are far more complaints if there is NO filtering!

Mark

I'm an Earthlink subscriber and if I send a HTML formatted message to an AoL subscriber I'll get back a 550 error.

Apparently AoL has received spam from an Earthlink domain so routinely blocks incoming traffic from Earthlink. (It seems to let plain text through, though.) Its a good bet that Verizon's doing the same thing. (Knowing how large companies work they've probably bought some anti-malware software from some slick salesperson, installed it and will be using it for ever because it would be a mistake to admit that its really crappy.)

Sigh.

Name supplied

As a matter of fact this is true. All our e-mail from the Czech Republic is blocked and it has been going on for at least 2 wks and they will not do anything about it till middle of January. I was told to get a hotmail account.

I too was told to send them list of e-mail addresses. I guess guilty till proved innocent. The only thing I can think about is sending them "bye-bye" notice. All our "Merry Christmas" and "Happy New Year" e-mails are gone. Merry Christmas to Verizon too.

You don't need to reply to this, I have no doubt that it doesn't work at least from some countries. I don't need another proof.

Name supplied

Wireless Irradiation

Go wireless? Are you nuts? That's radiation we're talking about. You can either catch testicular cancer from that, or if it's in the infrared light spectrum, probably some kind of skin cancer or what not. I'm sure we're going to see those warning labels put on everything wireless in a few years. From car key transmitters to WiFi relays. No, the wooden abacus. Surely thats the only safe way to compute these days

Happy new year
Søren Svendsen

AMD and Winchester probbo

AMD says there's no problem with Winchester chips. This "official" reply by AMD [INQ, dec 16th] shows that the person(s) at AMD did not read the thread mentioning the problems with some A64 chips at all - or knwowingly ignores issues people have with these CPUs.

The AMD official talks about "overclockers who are trying to get all components responding to a 25% overclock" - but the "issue" is rather a failure of certain batches of AMD A64 chips at rated, *stock* speeds.

It is beyond me how an AMD official can respond with "there is no problem with these chips" while, at the same time, some CPUs are unable to run certain applications like prime95 for stability testing at their rated speeds.

Problems are also reported with Knoppix Linux and other demanding tasks like kernel compilation under Linux.

Something is up here, at least with earlier batches of Winchester cores - and people could easily verify this just by swapping out their faulty CPUs for later batches and then NOT seeing any of these problems occur.

The irony now is that the author of Prime 95 denies anything wrong with the code of Prime95...as well as AMD denying "anything wrong with the AMD Winchester CPUs".

Just FIY, "Prime95" is a widely used tool used by computer enhusiasts to test a PC/CPU for stability. A processor should never fail these tests under normal recommended default settings and speeds. If it does it is defective/flawed since it is unable to "calculate right".

GW