Cisco Unified Wireless LAN Security Fundamentals
"New Vulnerability Allows Hackers to Penetrate Wireless Networks!" screams the headline in the newspaper or periodical. Perhaps the accompanying article describes some new theoretical vulnerability announced by a security research group that (surprise!) offers wireless LAN (WLAN) security consulting services. Or maybe it's a WLAN vendor that, quite naturally, not only "discovered" the new vulnerability but also offers the industry's "only" or "best" solution. Or maybe the accompanying article contains a sensationalistic description of how some "white hat" hacker demonstrated a new tool to exploit a WLAN or network attack vector at a security conference. Quite often, what's "new" is just a variant of what's old—a new exploit tool for a well-known vulnerability, for example. But then, every once in a while, articles of this ilk describe a significant new development that gravely impacts the industry.
Unfortunately, many journalists—even those writing for industry and technical publications—struggle to grasp even the fundamentals of WLAN technology, let alone the intricacies and complexities of WLAN security threats and their full ramifications on network design and implementation. It's shocking how often vulnerabilities common only in consumer WLAN implementations are applied in hysterical, sweeping generalizations to all wireless networks.
This is not to say that there aren't real security threats with WLAN networks; there definitely are some significant security challenges for WLAN network designers and operators. But the challenges are, for the most part, manageable when reality is filtered out of all the hype and the problem domain is well understood. Indeed, we often observe that the WLANs our customers deploy are more secure than their companion wired networks!
This chapter discusses the fundamentals of wireless LAN security in the context of the Cisco Unified Wireless Network (CUWN). An in-depth discussion and analysis of WLAN security can be its own book. In fact, there are already a number of excellent books available on the topic of WLAN security. Some favorites are listed in the references at the end of this chapter.
This chapter begins with an introduction of the security risks with WLAN technologies and continues with an explanation of technology building blocks that address and mitigate the risks.
When you are done reading this chapter, you should have sufficient background information on WLAN security. The security concepts discussed in this chapter are woven throughout the fabric of the CUWN. Indeed, one of the real benefits of the CUWN architecture is that it simplifies the design, deployment, and operations of security for your WLAN.
Understanding WLAN Security Challenges
You should know the vulnerability points of any network you are trying to secure and understand how the bad guys try to exploit them. How else do you separate the real from the hype and design sensible security policies and select the right WLAN security technologies?
This would be a good place for one of those hackneyed quotes about the importance of knowing your enemy from the likes of Sun Tzu's The Art of War. But all the good quotes we know of have already been used ad nauseum by other authors. So we'll spare you (and ourselves).
Instead, let's move right into discussing the security risks. The discussion that follows centers on the places where WLANs have security exposures as opposed to specific attacks and flaws. Basically this is because books have a long life, and by the time you read this, today's latest, greatest WLAN exploits might be old news. But the risk points remain the same. The risks discussed are as follows:
- Vulnerabilities inherent to the radio transmission medium
- Vulnerabilities inherent to the standards definitions
- Vulnerabilities inherent to mobility
- Readily available profiling and attack tools
- Misconfigured wireless devices and clients
- Rogue access points and devices
After concluding the following sections, you should have a good overview of the real risks associated with WLANs and be ready to take a closer look at the building blocks that address these vulnerabilities.
Vulnerabilities Inherent to the Radio Transmission Medium
WLANs have inherent vulnerabilities arising from the use of the airwaves and radio waves as the transmission medium. The two significant problem areas are
- Physical containment of transmissions
- Use of the unlicensed radio spectrum
The sections that follow look at these problem areas in greater detail.
Physical Containment Problem
With an Ethernet LAN, eavesdropping or attacking the network from the inside requires physical access to the network. Typically, an attacker must be able to connect a machine to a switchport in the network somewhere. Violating the network's security requires violating physical security.
This is not the case with WLANs. The basic physics of the transmission medium creates a physical containment problem. WLANs use radio signals over the air as the physical transmission medium. After a radio signal leaves its source, whether it is an access point or a wireless client, the signal travels through the air in many directions, and you have little or no control over the signal propagation.
Any listener with an antenna tuned to the right frequency and within range of the WLAN can "hear" the transmissions of both clients and access points. Skilled attackers know how to use high-gain directional antennas to profile and eavesdrop on WLAN networks from far away. But even relatively unskilled attackers can hear your WLAN pretty easily with simple tools.
If an attacker can hear transmissions in the unlicensed spectrum using readily available equipment from the WLAN coverage area, it's only logical that the attacker can also transmit into the WLAN coverage area relatively easily to cause big problems. Attackers might do this for one or more reasons. For example, the attacker might simply be trying to create a denial of service by using up available radio channel time. The attacker could also be trying to spoof a legitimate wireless device. It's not uncommon for an attacker to spoof a legitimate access point to try to trick wireless clients into connecting to the attacker.
Unlicensed Radio Spectrum Problem
The physical containment problem is exacerbated by the use of the unlicensed radio spectrum in both the 2.4-GHz and 5-GHz bands. Other types of wireless networks—for example, the cellular phone carrier networks—enjoy a certain amount of "security by obscurity" because they have a dedicated radio spectrum allotted to them by a regulatory agency. While that doesn't solve the physical containment problem, it makes it much harder for an eavesdropper or attacker because he has to obtain or build special equipment and tools to attack the network. This typically requires sophisticated knowledge and technical skills. It is also illegal.
On the other hand, with WLANs, attackers can use off-the-shelf equipment and open-source software attack tools. The skills and knowledge level required are moderate. Also, because the spectrum is unlicensed, the legal questions are much more abstruse.
WLANs can be susceptible to competition with non-802.11 devices that use the same radio channels in the unlicensed spectrum. From the 802.11 WLAN's perspective, this competition is considered noise, and if strong enough, can significantly degrade the network's performance. Common products that use some of the same spectrum as the WLAN are Bluetooth wireless devices, 2.4-GHz cordless phones, and microwave ovens. Legitimate devices don't represent a security problem per se, but they can affect WLAN availability, creating a de facto denial of service. Malicious attackers can use jammers to the same effect.
One of our favorite stories from Cisco sales lore comes from a customer bake-off between Cisco and a competitor for a large WLAN deal. The Cisco pilot was going very poorly and Cisco engineers were completely baffled because every failed test case in the pilot environment worked perfectly in Cisco labs. After many sleepless nights and much consternation, the mystery was solved. The competitor was camping out in a van outside the test environment with a doorless microwave oven jamming the airwaves during the pilot!
While this story is almost certainly apocryphal, it does illustrate how a legitimate product can be used nefariously and how the unlicensed spectrum makes the WLAN susceptible to RF jamming attacks.
Vulnerabilities Inherent to the Standards Definitions
The underlying IEEE 802.11 standards definitions have some inherent vulnerabilities, which fall into two categories:
- Authentication and encryption weaknesses
- Unauthenticated management and control frames
The sections that follow look at some of the details.
Authentication and Encryption Weaknesses
Put simply, authentication controls access to the network and networked resources by using techniques that identify who and which devices are allowed onto the network and those that are not. Encryption protects data frames in transit on the network, using cryptographic algorithms to obfuscate the frame content. When you consider the vulnerabilities inherent to the transmission medium, it's pretty obvious why both authentication and encryption are really important security concepts with WLANs.
The original IEEE 802.11 specification was released in 1997 and called out a mechanism for authentication and data privacy called Wired Equivalent Privacy, or WEP for short. This name is telling because it reflects the original goals of the standards designers—to provide a wireless data privacy mechanism roughly equivalent to what you get with a wired Ethernet network. In other words, it was supposed to be as hard to break WEP encryption as it is to violate an enterprise's physical security to gain access to the wired network. The WEP standard was designed to be a trade-off between "reasonably strong" security and implementation simplicity and exportability.
WEP is based on the shared-secret concept. Both end devices of a WLAN connection share a secret WEP key. The WEP key can be used to authenticate wireless devices; if a device has the secret WEP key, it must be authorized!
The WEP key is also used to encrypt data transmissions between each end of the WLAN connection. The original 1997 version of the 802.11 specification called out 40-bit WEP keys. In 1999, the specification allowed expanding the key length to 104 bits. These keys are statically configured on the devices that will use the WLAN.
As 802.11 WLAN technology started to take off, a lot of smart people in the cryptographic community started to take a good look at WEP as a security mechanism. In 2000 and 2001, several landmark papers were published detailing critical problems with WEP. If you're really interested, these papers are listed in the references at the end of this chapter, and they make for excellent reading to combat insomnia.
Not long after these papers were published, exploit tools appeared on the scene. These tools are now readily available on the Internet and are pretty easy to use, even for novices. So the most important thing to know about WEP is that it is irreversibly cracked and should never be used. It bears repeating: WEP is totally ineffective for data privacy because of cryptographic flaws; don't use it.
Recognizing that WEP was not the answer to WLAN security, the IEEE formed the 802.11i task group to come up with a robust security scheme for the future. The 802.11i task group's work was ratified in 2004.
While the 802.11i standard was in draft form, the Wi-Fi Alliance released its own requirements based on a subset of the 802.11i standard. The first iteration of these requirements was called Wi-Fi Protected Access (WPA). An update to these requirements is based on the complete, ratified 802.11i standard and is called Wi-Fi Protected Access Version 2 (WPAv2). The industry as a whole has moved toward 802.11i/WPAv2-based security, and that's where you should be too. Later in the chapter, you will learn more about WPAv2.
Unauthenticated Management Frames
Recall from the basics of 802.11 WLANs that there are three kinds of frames: control, management, and data frames. Discussions of WLAN security weaknesses are incomplete without noting that the 802.11 specification lacks an authentication mechanism for management frames.
The lack of authentication for management frames opens the door to a variety of denial of service (DoS) attacks. For example, an attacker runs a tool that spoofs disassociation and/or deauthentication management frames from the access point.
These DoS attacks can be run in conjunction with other attacks. For example, if you have a WLAN using Lightweight Extensible Authentication Protocol (LEAP) for authentication, an assailant could spoof deauthentication messages to all the users connected to an access point in the hopes of capturing username and password hash combinations when the client devices reauthenticate. If some username and password hash combinations get retrieved, an offline dictionary attack is used to crack as many passwords as possible. LEAP is covered later in the chapter, but the attack just described is why LEAP should not be used anymore. Coverage of LEAP has been included solely for historical and educational reasons.
Vulnerabilities Inherent to Mobility
The freedom offered by mobility is why we love wireless technologies; however, when it comes to WLANs, the same mobility that is the primary driver for adopting the technology also creates some security challenges.
One of the big challenges in an enterprise is figuring out how to handle roaming end users securely. Wireless clients regularly leave their association with one access point to reassociate with another access point. These wireless clients cannot just reassociate; they must be reauthenticated and generate new encryption keys. This means that the wireless client devices must carry some kind of security context with them so that the system can support fast reauthentication and rekeying if you want to avoid adversely affecting latency-sensitive applications like voice.
There are other problems inherent to mobility that are less related to technology and more related to end-user behavior. Suppose that you've deployed a secure WLAN in your enterprise, using the strongest authentication and encryption technologies available. You are confident in the strength of your WLAN security in your enterprise. But then, how do you secure the laptops of your road warriors when they connect to public hotspot WLAN networks in airports and coffee shops?
Consider what could happen when a senior executive in your enterprise uses her laptop to connect to an open network in an airport. There is probably some of your enterprise's important intellectual property and strategic information stored on that laptop in the form of documents, spreadsheets, PowerPoint presentations, and emails. You don't want an attacker compromising that valuable data. You also don't want that laptop catching a virus from another computer connected to the same wireless network. That laptop needs to be protected!
A similar problem arises with home WLAN networks. Home WLAN devices are very common these days. Usually, these are commodity devices from the local electronics megastore that don't always support the strongest security. Most end users aren't all that technical and don't pay much attention to security. These users get easily confused configuring security settings. Walk around any neighborhood with a WLAN sniffer and you will see that most home users don't give much attention to the physical containment problem either. Now when your enterprise users take their laptops home and connect to their home WLANs, how do you trust that the device is not vulnerable?
It can get worse too. A former neighbor of ours is a telecommuter, working as a marketing consultant for a large Fortune 500 company. His employer supplied him with a hardware Virtual Private Network (VPN) solution. Quite innocently, he decided it was a great idea to add a wireless access point so that he could enjoy working outside on nice days. And as you'd expect from a marketing consultant, his access point was configured with weak security. It never occurred to the neighbor that computers connected behind the hardware VPN client have a free ride onto his employer's corporate network through the VPN tunnel.
Misconfigured Wireless Devices and Clients
The previous section reviewed some of the security challenges inherent to mobility. Another issue that is often related is misconfigured wireless client devices. Client devices usually get misconfigured when users tinker with the client supplicant settings on their own, usually when they are trying to set up their home WLAN or connect to a public WLAN hotspot.
Wireless network devices, like access points, can get misconfigured too. We've been on a customer site where we (temporarily) crippled WLAN security during troubleshooting. This isn't necessarily a dumb thing to do; in this customer's case, we were troubleshooting issues with wireless client associations and were eliminating authentication and encryption as variables while working on RF issues. But it's pretty easy to forget to turn the security back on, especially after an all-night troubleshooting session!
Enterprise-class WLAN implementers can be presented with a dizzying array of configuration options, especially when it comes to some of the authentication and encryption settings. Even the most experienced network manager can make mistakes and inadvertently leave the network exposed in some way.
Rogue Access Points and Devices
Consider an enterprise with a wireless network deployment utilizing Extensible Authentication Protocol–Transport Layer Security (EAP-TLS) authentication and Counter Mode with Cipher Block Chaining Message Authentication Code Protocol (CCMP) for privacy. Don't worry if these concepts are foreign to you because they'll be explained shortly. Suffice it to say, for now, that this is a very strong authentication and encryption approach.
Now suppose though, that in some of the buildings, the access points are placed improperly so that some labs and conference rooms along the building periphery get poor radio coverage and users have difficulty connecting to the wireless network, and when they do connect, they experience very bad performance. Take it as axiomatic that people love the freedom of wireless mobility, so the poor end-user experience in the conference rooms and labs creates an unintended incentive for employees to deploy "rogue" access points.
Some employee, almost inevitably and quite innocently, will bring a cheap, commodity access point from the local electronics superstore into one of the conference rooms or labs with poor coverage, find a free Ethernet jack, and deploy an unauthorized, rogue access point, probably with weak security at best.
Clearly, this represents a catastrophic network security hole. The conference room or lab locations along the periphery of the building almost guarantee that the access point radio signals will be accessible from outside the building. Attackers frequently look for poorly secured WLANs to exploit, and it doesn't take much in the way of technical skill to find them. If and when an unauthorized user associates to the rogue, the user has free access to the enterprise network and can do all sorts of nefarious things.
In this example, the authorized wireless network is securely implemented with strong authentication and encryption. But while there is no glaring weakness with the official wireless network, there is a serious wireless security problem!
This example illustrates what Cisco has often called the "frustrated insider" rogue access point. These are rogue access points deployed by insiders out of frustration because of no wireless access or rotten WLAN performance.
There's an entire different class though of "malicious attacker" rogue access points and devices. These are rogue wireless devices implemented by the bad guys for the singular purpose of compromising your network. It's not hard to imagine a parasitic attacker tailgating an employee in your enterprise to bypass building security, then finding an available Ethernet jack and deploying a rogue access point that he can later exploit from outside the building.
There is also software readily available that can turn any computer with a wireless network interface card into a software-based access point. Attackers use these software-based access points to entice wireless clients to connect to them. After a wireless client connects, the attacker attempts to trick the wireless client into giving up valuable information, or else the attacker compromises the client device in some way. This attack vector is especially effective in public hotspot environments.
Readily Available Profiling and Attack Tools
So far, you've learned about the vulnerability characteristics of the radio transmission medium, vulnerabilities in the standards definitions, vulnerabilities introduced by mobility, and the challenge of rogue access points.
All the problems are exacerbated by the proliferation of profiling and attack tools on the Internet that exploit the basic vulnerabilities in WLANs. Many of these tools are very easy to get started and not hard to use. There are bootable Linux CDs that include all the latest tools and client card drivers that make running these attacks "chimp simple."