On Tue, 7 Sep 2010, Christopher Morrow wrote:
it used to be (~4-5 years ago) that the spammer code of 'voip service provider' was really 'we intend on raping proxies all over the planet' ... when you call them out on the random port traffic out of their pipe they point at their 'business' model that this is 'voip traffic, you know that rtp uses random ports, right?'
I haven't seen that excuse/justification from customers. What I did see recently that I have to admit was very slick was a customer who claimed they were going to be doing a bunch of remote "terminals" in stores VPN'd into their dedi servers and would be streaming video from the servers to the clients. This was of course 99% BS. There was VPN involved....they used the dedi servers as VPN endpoints for their spam servers that were hosted elsewhere. When we shut them down, there was absolutely nothing incriminating of spam operations on their servers...and all they had to do was sign up for service at another hosting company, setup the VPN server, change the IPs their spam servers VPN to, and they're back in business. When sales brought me their initial request, I really didn't believe it, but I didn't have good enough cause to reject it.
I used to have some quick/dirty instructions for how to verify that the traffic was in fact proxy traffic, something like: 1) log traffic from the soon-to-be-ex-customer (acl logs are fine) 2) pick an external 'top talker' 3) route that /32 to a host you control 4) run NC on the port that /32 is being contacted on 5) rejoice (and shut now ex-customer interface) when you see: "CONNECT smtp.xxxxx:25"
Seems like a lot of work when you could just setup a monitor session on their port and capture a few minutes of actual spam traffic as evidence just before shutting their port. ---------------------------------------------------------------------- Jon Lewis, MCP :) | I route Senior Network Engineer | therefore you are Atlantic Net | _________ http://www.lewis.org/~jlewis/pgp for PGP public key_________