Slashdot has a story about a research report form a Toronto-based consulting firm called IPEverywhere, which maps out "naked" access points in Toronto, implying that these nodes are all accidentally left open and are security risks. Here's my Slashdot post on the subject:
I just spoke with the COO of the IPEverywhere about this study, and confirmed that the methodology only established whether a node was running WEP (a "security measure" of dubious value).
That means that many of the "unsecured" nodes in this report may have had other means of securing themselves, from switch- or AP-based MAC filtering to captive portals such as NoCat. Moreover, the protocol for this study did not establish whether the open APs in question were handing out DHCP leases (or, indeed, whether they were connected to the Internet at all).
Finally, this study did not investigate in any depth whether the open APs were deliberately or accidentally left open. Many of us run open "community" networks around the world (I operate one in Toronto at King and Niagara, and three in San Francisco, two at 19th and Shotwell, and one on Sycamore near 17th and Mission). These networks are deliberately "unsecured" and are provided out of public-spiritedness, or even out of a political commitment to providing tools for anonymous speech on the Internet — anonymous speech being fundamental to democratic discourse.
Since WEP is such a poor "security" measure, the best practice for wireless users is to use SSH and/or SSL tunnels to secure sensitive traffic to a proxy (either remote or on your own network). In fact, if you're a promiscuous user of any network — conference centers, airport lounges, hotel rooms, schools, etc — you should assume that unless your messages are encrypted, they will be sniffed on the wire.
The primary "security" concern about open wireless seems to be that a "rogue" AP will be installed behind a firewall. The firewall, of course, is hardly sufficient in and of itself for securing a network. It's based on the presumption that everyone on one side of the firewall is trustworthy, and everyone on the other side is untrustworthy. We know, though, that this is a fallacy. Getting inside the firewall — either through physical intrusion (think of visitors to your office plugging into the the network to check mail) or virtually, by 0wning a box on the network with a trojan — is not difficult for a determined intruder. Meanwhile, the legitimate users of your network resources are often outside your firewall (mobile execs at a client site, for example) and thus not only walled off from the rest of the network, but also vulnerable to attack, since their machines' first line of defense is the firewall, which they are suddenly out of.
Security is hard. The proper place to draw your network perimiter isn't around your office, but around each machine. Personal firewalls, regular applications of security patches, good passwords and user education provide genuine security. Firewalls (and FUD about open APs) don't.
(via /.)
