networked inference

Yesterday we found license plate readers and want to learn more about them. At the end of the last post we had three IP addresses from a search of Shodan for “AutoVu” in the html field. AutoVu is the proprietary line of Genetec under which it sells a variety of automatic license plate recognition hardware and software. Two of those IP address gave response pages with LPRaaS in the title, for “License Plate Reader as as Service.” and one titled Login - AutoVu Occupancy Study, all on the Microsoft Azure cloud. The latter and one of the former IPs are in Boydton, Virginia, according to Shodan, and one, 23.101.179.17, is in Dallas.  Before digging more into these hosts I wanted to try some more searches on Shodan -- I went to a police magazine website’s product index and made a list from the ALPR page. I didn’t need it yet; searching Shodan for html:”lpr” and html:”alpr” was plenty -- it produced dozens of leads. This post is the first lead to pan out (I eliminated s...

This will be a post in three, well, probably three, parts, which I will probably post to Medium once done as a single thread. The eventual idea is to give a more complete picture of what I am trying to do -- and what I mean by networked inference. For now I want to proof a concept and get some of you along for the ride. Don't think of yourselves as bookish researchers, think CSI-Woke. When someone says a file or ‘page’ is on “the Internet,” they’re saying it’s on someone else’s computer where it can be accessed from another. The computers themselves must communicate according to a sort of heuristic, and “Google hacking” for research is really about understanding that vocabulary as well as that of the subject matter, how to find the lowest common denominator to make your task as fruitful and *automated* as possible. For this post I’m trying a cold search, “live,” if you will, in that I am typing these words and have only now conducted some preliminary information gathering (whic...