Over that weekend, 08/17/2008, the building that houses our office was struck by lightening. Except for 1 piece of equipment, there was no damage. However, that 1 piece of equipment, our T1 line, is essential to doing business. The other 11 people in that office are web developers and our ability to do our work depends on being connected to our company intranet and the internet at large.

So far actually getting Sprint to send someone out to look at it is a nightmare. They’re confident that the problem is on our end and we’re really confident that it isn’t. They just want to talk remotely to the “smart jack” and run “loop test” and say “it’s fine on our end”. Well come look at it in person from my end you jerks!

Lots of phone calls and still nobody is coming to take a look. Hell, I’m about ready to rip the damn box out of the wall, stop paying for the service I’m not getting and start paying somebody else, perhaps even the local cable company for a business connection. May not be T1 speed but then again I don’t have T1 speed now.

** Updated 2008.08.20 ** As of this morning we’re back on line. After hollering at the Sprint Customer Service rep they finally managed to send someone out. Turns out there were two problems one on their end and one on ours, but it was impossible to see ours until the alarm light on the T1 box was fixed.

After puttering around with the system for a little over a day I can honestly say that once I figured out how it worked, I ended up liking it. It’s really made a dent in the amount of daily spam that I’m receiving even while retaining a “catch-all” capability. What I did was the following:

1. Merged most of the domains that I manage into a common domain from Postini’s point of view so that mail coming from those domains is treated by a single set of Postini rules.
2. Created Postini based email addresses that matched the specific email addresses that were serving as “catch-all” mail boxes at the various domains.
3. For each mail box, established a list of Postini “user” aliases that I wanted to run the Postini spam filtering against. Addresses not matching the mail box name or the aliases go through the Postini system without filtering.
4. From the spam emails that were avoiding the Postini filter from step #3, I pulled out the most commonly used addresses and through my ISP established email forwarding rules that directed those addresses to a non-existent account causing emails using that address to bounce. In some specific cases I added the address as an approved “user” alias at Postini.
5. Kept repeating step 4 until the level of spam fell to near zero

Sounds like a lot of work, but in reality it wasn’t. The hardest part was wrapping my head around the way Postini handles catch-all addressing in the absence of a specific option. I had to just “trial and error” it until I figured it out. Once I understood the “rules” I quickly figured out how to leverage the Postini configuration and my ISP’s email forwarding rules to work in conjunction to achieve a better email filtering system. In essence I’m using the stuff that gets through Postini as a feedback loop to the email forwarding system. Right now step 4 is a manual process but I could easily make that programmatic.

At any rate, by the end of day 2 my email has gone from near 1000 pieces of spam per day down to a managable 100 or so and dropping every time I get new spam. Not bad for a few hours invested in figuring out the constraints and learning to work effectively within them.

Initially I though that one of my friends became part of a “bot net” and was unknowingly sending spam emails. Turns out it’s an internet-wide issue – read it yourself at Slashdot.

** update ** 08.07.2008 – they are still coming in at the rate of 200-400 per day