On 2/15/10 1:26 AM, CLIFFORD ILKAY wrote:

While scam emails are a nuisance, and perhaps apropos of nothing, a great deal of the attempted credit card fraud that we see originates in the good old United States of America.

We use geographic blocking simply to block unprofitable or nuisance traffic. For instance, one of the issues we had with our proxies were requests from Iran consuming huge amounts of bandwidth. While I am sympathetic to the fact that internet access is restricted there, advertisers who pay our bills could care less. They don't pay much if anything for impressions in Iran. Another country created problems with excessive downloads of large files, again consuming bandwidth at what seemed like all hours. We didn't want to rate limit everyone because the issue was really with one country, so we used a redirect to a different domain. That domain *only* had traffic from that one country, and requests were rate limited after a couple of megabytes. This had the effect that smaller files were easy to download but large files went very s-l-o-w-l-y. Users quickly adapted their usage of our service.

For credit card processing, we pre-screen with MaxMind's paid service. I think it costs $0.004 per request. It's the best four-tenths of a cent I can imagine paying. Many if not most of the frauds never even get to our processor to decline.

We use none of this in "server security". For what it's worth, we see break in attempts from all corners of the globe. The Western Hemisphere is well represented.