A rogue access point is a wireless access point that has either been installed on a secure company network without explicit authorization from a local network administrator,or has been created to allow a hacker to conduct a man-in-the-middle attack. Rogue access points of the first kind can pose a security threat to large organizations with many employees, because anyone with access to the premises can install (maliciously or non-maliciously) an inexpensive wireless router that can potentially allow access to a secure network to unauthorized parties. Rogue access points of the second kind target networks that do not employ mutual authentication (client-server server-client) and may be used in conjunction with a rogue RADIUS server, depending on security configuration of the target network.
To prevent the installation of rogue access points, organizations can install wireless intrusion prevention systems to monitor the radio spectrum for unauthorized access points. Presence of a large number of wireless access points can be sensed in airspace of a typical enterprise facility. These include managed access points in the secure network plus access points in the neighborhood. A wireless intrusion prevention system facilitates the job of auditing these access points on a continuous basis to learn whether there are any rogue access points among them.
In order to detect rogue access points, two conditions need to be tested:
The first of the above two conditions is easy to test - compare wireless MAC address (also called as BSSID) of the access point against the managed access point BSSID list. However, automated testing of the second condition can become challenging in the light of following factors: a) Need to cover different types of access point devices such as bridging, NAT (router), unencrypted wireless links, encrypted wireless links, different types of relations between wired and wireless MAC addresses of access points, and soft access points, b) necessity to determine access point connectivity with acceptable response time in large networks, and c) requirement to avoid both false positives and negatives which are described below.
False positive occurs when the wireless intrusion prevention system detects an access point not actually connected to the secure network as wired rogue. Frequent false positives result in wastage of administrative bandwidth spent in chasing them. Possibility of false positives also creates hindrance to enabling automated blocking of wired rogues due to the fear of blocking friendly neighborhood access point.
False negative occurs when the wireless intrusion prevention system fails to detect an access point actually connected to the secure network as wired rogue. False negatives result in security holes.
If an unauthorized access point is found connected to the secure network, it is the rogue access point of the first kind (also called as “wired rogue”). On the other hand, if the unauthorized access point is found not connected to the secure network, it is an external access points. Among the external access points, if any is found to be mischievous or potential risk (e.g., whose settings can attract or have already attracted secure network wireless clients), it is tagged as rogue access point of the second kind, which is often called an "evil twin".
Any AP other than authorized AP is rogue. Network connectivity of AP to enterprise network is not a criterion for rogue detection. Administrator will have to painstakingly manually separate out friendly neighbor APs. The manual inspection needs to be done on ongoing basis as new neighborhood APs pop up and old ones are reconfigured. If manual inspection is not promptly and regularly done, it creates security hole. Needless to say that automatic prevention of rogue APs cannot be turned on as administrator will have to first decide if a newly detected AP is on the network or just a friendly neighborhood AP.
AP’s connectivity to monitored enterprise network is still not a criterion for rogue detection, but filtering of friendly neighbor APs is possible based on preconfigured wireless only properties of neighborhood APs – such as SSID, MAC Vendor OUI and RSSI. Well, this logic only appears better than the first one. As a matter of fact however, it only gives false sense of sophistication for the reasons described below.
Security hole: There is nothing that necessitates the wireless only properties to be different between wired rogue AP and friendly neighbor AP. One could easily bring in AP whose SSID and vendor match one of your neighboring APs, put it on low transmit power so that it appears distant and connect it to your enterprise network. This situation can even occur in non-malicious case, if the employee brings in low power commodity AP with default SSID. The system will incorrectly classify it as friendly neighbor as its SSID and vendor match your preconfigured template for friendly neighbor and its RSSI is low enough.
Manual inspection: If legitimate neighbor APs change their wireless side settings or if new friendly neighbor APs are deployed, they will not fit the pre-configured template for friendly neighbors. So turning on automatic prevention is a risk as well as frequent manual inspection will be required.
AP’s connectivity to monitored enterprise network is essential criterion in AP classification as the rogue AP threat definition mandates; of course in addition to that the AP is not on the authorized AP list. Wired network connectivity of every AP visible in air is instantly, automatically and accurately determined by the system. If it is not on the authorized AP list and connected to the monitored network, it is rogue access point. If it is not on the authorized AP list and not connected to the monitored enterprise network, it is external (friendly neighbor) AP.
Automatic prevention can be safely turned on and there is no security lapse. No manual effort is required either at the beginning to configure any neighborhood AP properties templates or on ongoing basis as new APs come up and old ones change their properties.