CA Security Advisor Research Blog

After trials, US operator T-Mobile USA has released two new handsets that enable users to communicate either over its cellular infrastructure or over WiFi. The technology behind this new service called HotSpot@Home is Generic Access Network (GAN), formerly known as Unlicensed Mobile Access (UMA). In the presence of a 802.11b/g network, a UMA capable handset can obtain an IP address from the wireless Access Point (AP) and seamlessly handover an ongoing communication to the WiFi network. It is likely that the HotSpot@Home service (as well as other similar services by other cellular carriers) will become very popular. In fact, 2.5G cellular signals in the US –the most prevalent–are modulated at 900 MHz, 1800 MHz or 1900 MHz. The frequencies are not low enough to easily penetrate the thick walls present in urban environments. This means that customers will have very poor wireless coverage in a few scattered locations. The GAN protocol will help cellular phone operators reach more customers through virtual coverage.

The WiFi revolution is in full swing and is about to rock the cellular phone world the same way it did with laptops. WiFi has become so ubiquitous that in 2007, it is unusual for computer vendors to sell a laptop without it. Yet, connecting to a network via WiFi requires that one use more precautions than when connecting to a network with a wire. Let's take a look at the security risks involved with using WiFi.

Start with the obvious: communicating over an unencrypted wireless network is like shouting into a megaphone in Times Square and thinking that nobody will hear you: it’s just not going to happen. Hackers prey on open wireless networks (networks that do not require an encryption key to join). Packet sniffers set in promiscuous mode can intercept every single packet transmitted over the network while disguising their activity. If you care about your privacy, never transmit sensitive data over an open network.

Now that I’ve tried to discourage you from ever joining an unencrypted wireless network, you will hopefully connect to an encrypted network. There are two main wireless encryption schemes. WiFi was introduced to the general public with WEP (Wired Equivalent Privacy) encryption. WEP uses 40-bit or 104-bit keys. While WEP keys were considered strong enough just a few years ago, in 2005 the FBI showed how to crack a 64-bit WEP key in 3 minutes using commonly available tools. It is therefore best to rely on wireless networks that encrypt traffic with the latest encryption methods, namely WPA (WiFi Protected Access) and WPA-2. Keep in mind that WPA and WPA-2 can be cracked too...it just requires more time but it can be done.

WiFi is susceptible to attacks that do not apply to wired networks. Moreover, the strongest encryption key cannot thwart the following attacks:

- Passive Password Cracking: Hackers can record the packets that wireless Access Points (AP) send to their clients. Wireless networks in crowded urban area are more exposed to this kind of attack. This method of cracking a passkey is slow since the hacker needs to obtain many packets….but your neighbor could potentially perform this type of cracking for months without being detected.

- De-Authentication Attacks: During the authentication process, a few packets are exchanged between a host and an AP. Hackers are particularly interested two of those packets: the encrypted challenge text sent to the client and the client response to that challenge text which contains the challenge text encrypted with a shared key. Both of these packets are encrypted using the shared key and can be used to crack the key. Once the key is cracked, a hacker can spoof AP MAC address and send a de-authentication request to the client. Software on the client will usually attempt to automatically reconnect to the AP but the attacker can keep on sending deauth frames to keep the client off the network. Moreover, these frames are not seen by the real AP since they are sent by the attacker directly to the client.

- Injection Attacks: A password can be cracked on networks with little traffic. In fact, an attacker just needs to listen for an ARP frame and replay it to the AP after changing the frame’s origin address. The AP will respond to the sender of the ARP frame with many encrypted frames…which is exactly what the attacker needs in order to quickly crack the encryption scheme

- Man-In-The-Middle Attacks: Hackers can configure a rogue AP to imitate a legitimate AP. Once the client is connected to the rogue AP, the hacker can perform any attack that involves modifying the packet stream. Emails can be read, phishing attacks can be implemented etc...

These attacks can be targeted to more than just laptop computers. During a handover from cellular to WiFi, a handset acquires an IP address and thus becomes vulnerable to the attacks that were mentioned here. Cellular phones are increasingly being referred to as “smartphones” or “personal communicators” because they pack many features that can be found on full-size computers: Java Virtual Machine, web browser, email client….To compound the security risk that WiFi poses, Wireless Intrusion Protection Systems are still relatively new and are not designed to effectively protect networks from the type of attacks mentioned here. Be mindful of the risks you run when connecting to a WiFi network if you do not want to share your data with the world, a la Paris Hilton.