Myth vs. reality: Wireless SSIDs

Do you ever wonder sometimes how it is that some ideas just won't die? Like the thought that not broadcasting your wireless network's SSID will somehow make you more secure? This is a myth that needs to be forcibly dragged out behind the woodshed, strangled until it wheezes its last labored breath, then shot several times for good measure.

Folks, there are fundamental differences between names, which are public claims of identities, and authenticators, which are secrets used to prove identities, and I've written extensively about this before. An SSID is a network name, not -- I repeat, not -- a password. A wireless network has an SSID to distinguish it from other wireless networks in the vicinity. The SSID was never designed to be hidden, and therefore won't provide your network with any kind of protection if you try to hide it. It's a violation of the 802.11 specification to keep your SSID hidden; the 802.11i specification amendment (which defines WPA2, discussed later) even states that a computer can refuse to communicate with an access point that doesn't broadcast its SSID. And, even if you think your SSID is hidden, it really isn't. Let me explain.

All 802.11 wireless networks, regardless of the kind of operating system or encryption you might use, also emit unencrypted frames at times. One kind of unencrypted frame is an association frame. This is what a client computer, or "supplicant" in the 802.11 protocol vernacular, emits when it wants to join a wireless network. Contained within the frame, in clear text of course (since the frame is unencrypted), is the SSID of the network the supplicant wants to join.

Both Windows XP and Vista work best when your access points broadcast their SSIDs. XP really doesn't behave well at all with nonbroadcasting SSIDs. Vista has some added smarts to improve this a bit. Normally, Vista continually sends probe requests for nonbroadcasting networks. These probes are similar to unencrypted 802.11 association frames, and will generate clear-text responses from the access points if a nonbroadcasting network is present. You can reduce, but not entirely eliminate, these probes by configuring the wireless client to probe only for automatically-connected nonbroadcasting networks.

Both these behaviors make it very easy for an attacker to discover your SSID. The bad guy, perhaps a contractor or a guest in your facility, could run one of many wireless sniffer programs and simply capture the hundreds of association frames or probes that litter your air. No amount of "hiding" configured in your access points can prevent this kind of traffic interception.

So there you have it, simple SSID discovery. The old axiom remains true: security by obscurity is no security at all. Hiding an SSID will not hide a wireless network, so ignore any such advice -- and it's amazing how often I continue to see this. By the way, also ignore any advice that says to use MAC address filtering. It's amazingly trivial to spoof the MAC address of an allowed supplicant -- simply sniff the traffic, look at the MAC addresses, and use the neat little SMAC utility to change your MAC to one that's permitted.

Nonbroadcasting networks are not secure networks. The right way to secure a wireless network is to use protocols that are designed specifically to address wireless network threats. If you're still using WEP, either static or dynamic, I encourage you to move to WPA2 as soon as possible. For those of you at home running XP and have kept it updated, or if you're running Vista, then, you simply need to enable WPA2. We've got some additional guidance for home/small offices and for enterprise networks with certificate services or without. If you have hardware that's more than two years old and you can't upgrade it, check to see whether it supports WPA (an interim specification released before WPA2 was ratified). Both WPA and WPA2 are built on sound cryptographic principles, they're proven in the field, and they'll keep the bad guys out -- even when you're broadcasting your SSID to the world.