Performance on various domains - ANTICHEF.COM

[tousavelo]

Hello,

I have experienced delivery delays of one (or possibly more) hours with ANTICHEF.COM (my so-far preferred domain).

In a few comparison tests I made, I never experienced long delays with spamgourmet.com. However, antichef.com was sometimes pretty slow (most striking example: 2s for spamgourmet.com; 52 min. for antichef.com ).

Are some domains consistently faster than others ?
Has antichef.com suffered of some problem lately ?

Thank you very much in advance
Best regards
O.

PS: I could just switch to spamgourmet.com but I have spent lots of hours on updating my profiles on various web site with ANTICHEF.com ...

PPS: The design of spamgourmet is absolutely brilliant. Nice piece of art.

[dl97]

Am experiencing hour+ delay on a-bc.net, but spamgourmet.com test emails arrive in ~10 mins (not great, but better than an hour). Any ETA on a fix?

Thanks for the great service!

[info]

thanks!

There really shouldn't be any difference on the domains that are published -- they're all set up the same way with the same hardware.

Some unpublished ones have to take an extra trip before they arrive, but not the ones you all are talking about.

So - weird -- how contemporaneously were the tests sent?

[tousavelo]

EDIT: Order of the list of tests changed to chronological order. Clarification.
***************************************

Hello,

Sorry for the delay.

The original tests were run on 5 oct and 7 oct.
Three more comparison tests executed yesterday.

Target (don't watch the suffix):
testperf2.XXXXX@spamgourmet.com (testperf2 message 6 of 20).eml
testperf1.XXXXX@antichef.com (testperf1 message 6 of 20).eml

Time:
TEST5: Tue, 9 Oct 2007 09:44:36 -0700
TEST6: Tue, 9 Oct 2007 10:00:02 -0700
TEST7: Tue, 9 Oct 2007 11:31:20 -0700


Method: Two emails, each sent to one of the two domains, in the same second (or so). Duration computed as last timestamp before yours (hotmail), first timestamp after yours (small ISP in Belgium).

Duration:
Antichef.com:
7h
24min
10min

Spamgourmet:
less than zero (indicates a bad time setting in a mail servers after yours)
10min
10min

So, we have one example with similar duration, one twice as long, and another with a "out of bound" duration. When there was a noticeable differences, antichef was the slower domain.

The funny thing was that, in this batch, the first message sent to antichef arrived last, hours after the two other messages sent afterwards to antichef.

I'm wondering how all other domains compare. Is spamgourmet the fastest one or do you have an ultra-high-speed one .

Hope it helps and you can take some action.

I could email you the email files as .eml if you like.

Best regards
O.

[tousavelo]

EDIT: A-BC emails were missing due to bad email adress...
*******************************************

More tests with spamgourmet.com (SG), antichef.com

TEST08: SG: 11Min; ANTICHEF: still missing, 1hour+
TEST10: SG: 19Min; ANTICHEF: 1h15min
TEST11: SG: 10Min; ANTICHEF: 25min

Again, SG is faster. The other is often much slower.

PS: Time of tests, today, Paris time (first timestamp is departure time, last timestamp is arrival time):

‎TEST08 22:20 testperf2.XXXXXX@spamgourmet.com (testperf2: message 7 of 20) - 22:31
‎TEST10 22:23 testperf2.XXXXXX@spamgourmet.com (testperf2: message 8 of 20) - 22:42
‎TEST11 22:45 testperf2.XXXXXX@spamgourmet.com (testperf2: message 9 of 20) - 22:55

[tousavelo]

Hello again,

BTW, the 3 messages to spamgourmet.com have been received in the order they were sent. The 3 messages to antichef have been received in the exact opposite order they were sent :

Sent: TEST08 then TEST10 then TEST11.
For antichef: Received TEST11, then TEST10, then TEST08 (actually TEST08 is still missing but let's assume I will ever get it).

O.

[tousavelo]

Hello,

I have run another somparison test but I can't yet fully analyze the results. (I'm at work right now and my webmail interface only displays the "sent" timestamp)
My first feeling is that the results are normal (no major discrepancy between domains, just the expected variations from one email to the other).

I'll post more info. when available.

BTW: In this last setup, I used a new SG account and the route was large belgian ISP to SG to hotmail. In previous tests, route was hotmail to SG to small belgian ISP to gmail. (Note that I have used the small belgian ISP and gmail for quite some time and never noticed unexpected delays when SG was not in the picture.)


Best regards
O.

[SysKoll]

Thank you for your tests!

Hotmail is pretty unreliable. We cannot really be sure if there is a fault somewhere along the way when Hotmail is involved, because they routinely discard email, both outgoing and incoming. "Hotmail discards emails" returns 354,000 results on Google. Here is one: http://www.news.com/2009-1023-251171.html?legacy=cnet
Also see this thread: http://archives.neohapsis.com/archives/ ... /1097.html

So please do not limit your tests to Hotmail, try to perform the same tests with a proven email provider.

I don't seem to have problems with gmail.

[tousavelo]

Hello,

Here is more info about the tests I made.

To save you some time, here is the bottom line: all domains seem indeed equal. But each email may be very different, regardless of the domain, with a delivery time ranging from instant to several hours.

The first round of tests (reported in previous posts) did involve an additional layer of forwarding (beside spamgourmet). I have discarded those observations. The results reported bellow were sent from a major ISP in Belgium, to GMAIL.

AVG is the average, MED is the median (50% of emails were delivered faster), PCT(80%) is percentile 80%: 80% of mails were delivered faster, idem for PCT(90%)

For antichef (based on 20 emails):
MAX: 1h29, AVG: 0h19, MED: 0h04, PCT(80%): 0h32, PCT(90%): 1h01
For spamgourmet (based on 20 emails):
MAX: 1h29, AVG: 0h17, MED: 0h05, PCT(80%): 0h53, PCT(90%): 1h04
For other published domains (based on 70 emails):
MAX: 2h12, AVG: 0h17, MED: 0h07, PCT(80%): 0h18, PCT(90%): 0h55

[SysKoll]

Great job, thanks! The variability of delivery time is itself due to the random nature of spambots. Their activity is not constant. This means that the traffic on our machine and on the SMTP servers for the forwarding email addresses will vary greatly across the day.