Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 1

04/18/08

Running head: A BRIEF GUIDE TO SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR

A Brief Guide to Securing Wireless Networks: Closing the Back Door

in APA Style

Aaron Sawyer

East Carolina University

Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 2
04/18/08

ABSTRACT

Wireless computer networks (WLANs) are being used in greater numbers now than ever before. Due to

the unprecedented mobility, convenience, and flexibility that the technology offers, homes, businesses,

and enterprises are deploying wireless networks in increasing numbers. Unfortunately, wireless

networks may also act as an open door in the network security perimeter, allowing nearby attackers to

eavesdrop on confidential communications, break into the wired network, or cause other serious

problems. While the term "wireless security" has become an oxymoron for some in the face of recent

wireless attacks, wireless networks can be made secure. This paper aims to give a brief history of

wireless security, educate the WLAN administrator about the dangerous vulnerabilities of wireless

networking, and make practical suggestions about how to close the door on potential intruders. Even

though this paper focuses on the basic service and extended service set wireless architectures, many of

the concepts can be applied to any wireless network.

Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 3
04/18/08

TABLE OF CONTENTS

I. Introduction Page 4

II. Brave New World – The dangers of wireless networks Page 7

III. History of Wireless Security

A. First Steps in Wireless Security – Wired Equivalent Privacy (WEP) Page 10

B. It seemed like a good idea at the time – The Vulnerabilities of WEP Page 14

C. The Dam Burst Page 17

D. MAC filtering and Closed Networks have their own Issues Page 20

E. Wi-Fi's WPA to the Rescue Page 21

F. IEEE 802.11i & WPA2 Finally Arrive Page 25

IV. Common Wireless Attacks Page 30

V. Wireless Networks used Today Page 36

VI. Roads to Security – Commonsense Guidelines Page 40

VII. The Battle Continues Page 49

VIII. Works Cited Page 50

Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 4
04/18/08

I. INTRODUCTION - WIRELESS LANS – THE BACK DOOR

In July of 2005, outside of a Marshals store near St. Paul, Minn, a team of unknown attackers used a

laptop computer along with what is believed to have been a telescopic antenna to lock onto, monitor,

and eavesdrop on a wireless local area network (WLAN) used inside the store. After discovering that

weak WEP (Wired Equivalent Privacy) encryption was being used, the team cracked the encryption

scheme and began deciphering the wireless communications. As hand-held price checking devices,

cash registers, and company computers communicated with one another, the attackers recorded their

activities. The attackers also captured the usernames and passwords of employees as they logged in

(Pereira, 2007).

After gleaning enough information from the wireless network traffic, the attackers focused their attacks

on Marshal's parent company, TJX. Using the data that they had obtained, the group was able to break

into TJX's central database, retrieving sensitive customer information, including Social Security

numbers, driver's license numbers, military identification numbers, and credit card numbers. The stolen

information was soon available for sale by the attackers on password-protected Websites for identity

thieves (Pereira, 2007).

Signs of the heist first surfaced in November of 2005, when bogus credit card purchases began to

appear. It wasn't until December 18 that an auditor discovered oddities in card data, indicating that

something strange had taken place. Investigators were unable to catch the attackers since they used the

IP addresses of private individuals and public locations. Forrester Research estimates that TJX's

expenses from the breach could exceed $1 billion in five years. These include expenses from
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 5
04/18/08

consultants, upgrades, attorney fees, and increased marketing, but not for possible lawsuit liabilities

(Pereira, 2007).

The heist became known as the largest theft of credit-card numbers in history, with at least 45.7 million

credit- and debit-card numbers being stolen. In addition, some 451,000 customers had private

information stolen. Although the identity of the attackers is still unknown, security experts believe that

the group may have been Romanian hackers or members of Russian crime groups. The attack has

striking similarities to previous attacks perpetrated by these gangs; they are known for scoping out

weak targets and methodically breaking into networks. Investigators blamed the success of their attacks

in part on the weak wireless encryption scheme used (WEP). An auditor also later found that firewalls

and data encryption were not implemented on a number of computers connected to the wireless

network (Pereira, 2007).

While attacks of this scope and depth are relatively rare, the TJX case highlights the threats that

computer networks face in the new wireless world. Marshals is not alone in its continued use of weak

wireless security schemes. A study conducted by AirDefense Inc. in November of 2007 found that

twenty-five percent of the wireless business networks scanned in retail outlets and malls around the

world didn't use encryption of any kind. The same study found that another twenty-five percent of

networks were using insecure WEP encryption, and that twelve percent used the name of the store as

the SSID (Service Set Identification), giving potential attackers an instant “map” of the store. Other

networks were found to be using access points (APs) with default passwords that could be easily

broken into. In all, the study concluded that 85 percent of the wireless devices discovered would have

been relatively easily to compromise or to eavesdrop on (AirDefense, 2007; Wilson, 2007).

Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 6
04/18/08

Wireless security is a widespread problem. Often, wireless networks open a back door in an otherwise

secure computer network. This back door can form the weak point in the security perimeter that

attackers are looking for. For homes, small businesses, and large enterprises, implementing effective

wireless security is an absolute necessity. This paper will seek three major goals:

1) To give an historical background of wireless security,

2) To discuss the how each technology fairs against common attacks, and

3) To provide a list of ready-to-use wireless security practices that an organization can implement

There are three major wireless network architectures implemented today: the basic service set, the

extended service set, and the independent service set. The basic service set (infrastructure mode) is a

collection of wireless devices that are served by a single access point (AP). This configuration is used

in small wireless network implementations. The extended service set is composed of two or more basic

service set networks. It is able to cover a much broader area and can offer roaming services for wireless

clients. This is frequently used in large wireless network implementations. Finally, the independent

service set (ad-hoc mode) does not utilize an access point, but instead is completely peer-to-peer

(Ciampa, 2006, pp. 154-156). Because most organizational wireless implementations utilize the basic

service set and extended service set architectures, this paper will attempt to address them primarily, but

the principles discussed could be applied to other topologies.

Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 7
04/18/08

II.THE DANGERS OF WIRELESS NETWORKS – BRAVE NEW WORLD

The fundamental characteristic of wireless networking is the ability to connect to a computer network

without physical contact. Wireless networks broadcast data over publicly accessible radio frequencies

instead of transmitting over private wires. In the past, businesses only dealt with wired networks. Such

networks had few entry points and were generally harder to infiltrate from the outside. Attackers would

have to connect to the physical LAN topology to infiltrate the network. Back-door attacks which

penetrated a network from the inside were rare due to the feasibility of such an attack. See Figure-1 for

a simple example of a small business wired network.

Firewall

Workstations

Parking Lot Proxy Server

Street
Switch
Windows

File Server

Windows

Street

Figure-1. A typical small business LAN.

Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 8
04/18/08

With the introduction of wireless networks, the rules changed. No longer does an attacker have to

connect to the physical network. Instead, an attacker can break into the network from across the

parking lot. Today, attackers can scope out potentially vulnerable wireless networks from a moving

vehicle, a popular practice known as wardriving.

Wireless networks open new doors into networks that previously did not exist, and can bring with them

catastrophic consequences. As more and more small businesses and large enterprises implement

wireless networks for increased productivity, mobility, and freedom, they often overlook the fact that

they have just opened a gaping hole in their network's defenses. Refer to Figure-2 for the small

business network with a wireless connection. Notice that due to the wireless network broadcasting

signals in all directions, the network is now accessible not only from within the intended building area,

but also in the business's parking lot and even across the street. The wireless traffic travels even farther

through unobstructed windows. This simple diagram does not show the additional wireless traffic

emitted by clients.
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 9
04/18/08

Firewall

Workstations

Parking Lot Proxy Server

Street
Switch
Windows

File Server

Access Point

Windows

Street

Potential Attacker

Figure-2. A typical small business LAN with a wireless connection.

Due to these new vulnerabilities, wireless networks must be secured to protect the wireless data in

transit as well as the internal wired network. As the TJX case demonstrates, wireless networks can form

the stepping stone for an intrusion into the central network. The broadcast nature of wireless LANs

makes security a particularly important subject.

Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 10
04/18/08

III.THE HISTORY OF WIRELESS SECURITY

A. FIRST STEPS IN WIRELESS SECURITY – WIRED EQUIVALENT PRIVACY (WEP)

Even in its infancy, wireless network technology implemented a degree of security and privacy. The

standard governing wireless networks is known as IEEE 802.11. As the original Institute of Electrical

and Electronics Engineers (IEEE) 802.11 standard was being developed, board members were aware of

the new security challenges that broadcast communications brought with them. IEEE members saw that

wireless networks needed to be protected from eavesdropping, unauthorized access, and data

tampering. The team reasoned that because the confidentiality, access control, and data integrity of

information is under increased threat over public airwaves, some mechanism was needed to bolster

security to the level found in a wired network, at a minimum (Wong, 2003).

To provide data confidentiality in a wireless network, the IEEE decided that an encryption scheme was

necessary to encrypt, or “scramble” the data as it traveled from one end point to another. If messages

are encrypted before they hit the airwaves, they should be unreadable by unauthorized wireless users.

Thus, encryption was proposed to prevent casual eavesdropping and provide data confidentiality

(Brown, 2003).

The scheme developed by the IEEE group was WEP, Wired Equivalent Privacy. The system was meant

to provide the “equivalent” amount of security thought to be present in a wired network. WEP was

designed to implement the RC4 (Rivest Cipher 4) stream cipher, combining a 40-bit key with a 24-bit

(theoretically) random number known as an Initialization Vector (IV) to encrypt messages (Wong,

2003). A stream cipher is an algorithm which replaces each character with another coded character.
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 11
04/18/08

Thus, each character is individually encoded (Ciampa., 2006, p. 270). Refer to Figure-3 for a

description of the WEP encryption process.

Step 1
Plaintext CRC = Plaintext ICV
Step 2 Step 3
Initialization
Vector (IV)
+ Secret Key PRNG = Keystream

Step 4 XOR

Step 5 IV Ciphertext

Figure-3. The WEP Encryption Process.

Source: (Ciampa, 2006, p. 269)

In step 1, the unencrypted data (plaintext) to be encrypted is analyzed and a cyclic redundancy check

(CRC) value is calculated. The CRC creates a unique checksum based on the plaintext message which

verifies the contents of the message. WEP refers to this as the Integrity Check Value (ICV) and it is

appended to the end of the plaintext message. In step 2, the secret WEP key is combined with a 24-bit

value, known as the Initialization Vector (IV). The IV is serves as a random value which changes with

each packet. It helps to ensure that a random number will be generated in the next step. Without this

random value, the keystream could be the same every time, and the encryption could be cracked. In

step 3, the shared WEP key and the IV are entered into a pseudo-random number generator (PRNG) to

create a random number: the keystream. The keystream is a series of 1's and 0's equal in size to the

plaintext message and ICV. In step 4, the keystream, along with the plaintext message and ICV, are

combined into a single message through an XOR operation. The result is ciphertext – the encrypted

message. Finally in step 5, the IV value is added (without encryption) to the beginning of the ciphertext
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 12
04/18/08

because the message cannot be deciphered without it. At this point, the IV and ciphertext can be

transmitted. A receiver can use its copy of the shared WEP key to backtrack through the process to

produce the original message (Ciampa, 2006, p. 269).

To provide wireless authentication and access control, the IEEE team made two authentication types

available:

1. Open system authentication

2. Shared key authentication

Open system authentication is not truly authentication at all. All a wireless user needs is the SSID, or

name of the network to join. If the user knows the name of the network, he can join the network. Open

system authentication is therefore an “open” system. WEP could still be used to encrypt messages, but

it is not used to authenticate new devices wanting to join the network (Ciampa, 2006, pp. 270-271;

Microsoft Technet, 2006)

In shared key authentication, the WEP key is used to provide data confidentiality and to restrict

network access. If WEP-enabled wireless devices receive packets without proper WEP encryption, the

packets can be ignored, thus theoretically limiting the secured wireless network to only those devices

which have the WEP key (Brown, 2003).

Lastly, to provide data integrity, the 802.11 standard included an integrity checksum value, computed

by both receiver and transmitter. If the checksum does not match, the packet is discarded. Thus, if any
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 13
04/18/08

packet happened to be manipulated, it would theoretically be ignored (Brown, 2003).

Outside of the IEEE 802.11 standard, two other notable access control mechanisms were soon

developed by wireless device vendors:

1. MAC Filtering

2. Closed Networks

To help restrict unauthorized wireless access, vendors offered a MAC (Media Access Control) address

filtering option for wireless access points. The MAC address is a unique number made up of 12-

hexadecimal digits which are “burned” into all network interfaces. Because it is unique to each

interface, an administrator can theoretically deny access to any wireless device not specifically allowed

in a preconfigured list. Although this became a useful option in many small networks, it brought with it

its share of disadvantages, most notable of which is the trouble of maintaining a list of static addresses.

The “closed network” concept is a proprietary access control method created by Lucent. In this scheme,

the SSID of the network is kept a secret, and only those stations knowing the SSID may join the

network. Thus, if the SSID is indeed secret, only authorized devices will be able to join the network

(Arbaugh, Shankar, Wan, & Zhang, 2002). The option to disable SSID broadcasts has since been

offered by many access point vendors (Ars Technica, 2002, p. 3).

The original wireless standard, which included WEP, became known as IEEE 802.11 and was adopted

in 1997 (Wi-Fi Alliance, 2003). Unfortunately, the WEP encryption system was not subjected to
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 14
04/18/08

significant peer review before release, and it wasn't long until serious vulnerabilities were found

(Wong, 2003).

B. IT SEEMED LIKE A GOOD IDEA AT THE TIME... - THE VULNERABILITIES OF WEP

WEP seemed like a good idea at the time. At a time when encryption schemes were under tough

scrutiny due to export restrictions and few users had gone wireless, revealing few vulnerabilities, WEP

seemed like just the thing to make wireless networks secure, or at least as secure as their wired

counterparts.

In 2001, WEP's first vulnerabilities were revealed. Two major papers by researchers at Berkley and the

University of Maryland (UMD) found many vulnerabilities, including:

● Key reuse or “keystream attacks” due to “collisions”

● Weak message authentication

● Weak 802.11 access control mechanisms, including WEP cryptographic authentication

● Poor key management (the use of a single key by many users)

The Berkley team pointed out that the initialization vector (IV) used to make a random keystream for

packets is too small. Because the IV is just 24 bits long, it provides only 16,777,216 possible values,

and it will eventually repeat itself, creating an IV “collision.” Attackers could capture traffic on the

network, wait for an IV collision to occur, and then use the identical keystream to help crack the
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 15
04/18/08

keystream (a keystream attack). With an identical IV, if the plaintext of one of the packets can be

discovered, the keystream can be deduced (Borisov, Goldberg, & Wagner, 2001).

There are many possible ways to get the plaintext of the message. One can predict the IP headers used,

perform a dictionary attack (guess), or even attempt to inject traffic across the network. This could be

done by sending an email to a wireless user with known contents and waiting for the user to open it. At

any rate, once a keystream has been isolated, messages encrypted with it can be deciphered. If an

attacker can build a table of all of the keystreams used in the WEP network, the WEP key isn't even

needed to decipher communications; the keystreams can be used (Borisov, Goldberg, & Wagner,

2001).

The Berkley team also found that due to weak message authentication, encrypted packets could be

tampered with by having their ICV values modified. This could be accomplished by flipping the correct

bit in the packet, and is possible because RC4 is a stream cipher, encrypting each character

individually. Packets could then be injected back into the WLAN by using a previously used IV.

Packets with reused IVs are perfectly acceptable in WEP (Borisov, Goldberg, & Wagner, 2001).

A very important finding of the Berkley team was that shared key authentication is inherently insecure.

In shared key authentication, the device attempting to join the WLAN first requests to be authenticated

by the AP. The AP then sends a challenge back, which is a random 128-byte fixed length string. The

device requesting to be authenticated must respond by encrypting the challenge text with its WEP key

and sending it back. If the response is validated by the AP, then the requesting device is authenticated

and may join the network (Borisov, Goldberg, & Wagner, 2001).
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 16
04/18/08

The problem with the challenge-response method is that the challenge is sent in plaintext. An observer

views the unencrypted data and then views that same data encrypted. By passive observation, an

attacker can determine the keystream used and can use it to encrypt further messages and decrypt

messages using that same IV and keystream. Furthermore, because WEP allows packets to be injected

without the key (all one needs is the keystream and IV), an attacker can request to be authenticated and

respond to the challenge text with the same keystream and IV that was observed. Since the length of

challenge is always 128-bytes, the response will be valid. Thus, an attacker can become authenticated

in a shared key authentication network by simple observation (Borisov, Goldberg, & Wagner, 2001).

Even so, the attacker would still not be able to use the WEP-encrypted network without further attacks

due to the presence of encryption (Arbaugh, Shankar, Wan, & Zhang, 2002).

Because the challenge-response method used in shared key authentication is so flawed, shared-key

authentication is now considered to be worse than open authentication with WEP. In fact, Microsoft

now recommends that organizations stuck with WEP disable shared-key authentication and use open-

authentication with WEP data encryption enabled (Microsoft Technet, 2006).

Another issue arising out of weak authentication is the man-in-the-middle attack. Because

authentication is one-way, with the client being authenticated but not the AP, nothing is stopping an

attacker from setting up a fake AP (a rogue AP) and acting as a middle man for other users, passing

their data along to the legitimate AP, but also recording and/or modifying their traffic. This is possible

due to the cryptographic weaknesses of WEP and a lack of two-way authentication (Brown, 2003).
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 17
04/18/08

The Berkley researchers also noted that key management is often poorly implemented, and most

organizations use the same key for all wireless users. WEP attacks are easier when the same key is

used. Because of the hassle of reconfiguring the WEP key on every wireless client device, the WEP key

is largely static (Arbaugh, Shankar, Wan, & Zhang, 2002).

These vulnerabilities formed substantial cracks in the security of WEP. The popular consensus was,

however, that these flaws could be patched. Long-key length versions of WEP were created and

released, but they did not address WEP's weaknesses. Deceptive advertising by vendors didn't help

matters either. Vendors who advertised 64-bit WEP or 128-bit WEP encryption mislead consumers. 64-

bit WEP is actually composed of a 40-bit key and a 24-bit IV. Likewise, 128-bit WEP is actually a 104-

bit WEP key plus the IV. Even with longer encryption keys, WEP's problems remained.

C. THE DAM BURST

The eventual “bursting of the dam” came in 2001 when Scott Fluhrer, Itsik Mantin, and Adi Shamir

unveiled enormous flaws in the heart of WEP; in the Key Scheduling Algorithm (KSA) (Gast, 2002).

Fluhrer, Mantin, and Shamir found that RC4, “the most widely used stream cipher in software

applications,” had two significant weaknesses. First, it was found to have an “invariance weakness” in

which a small part of the key determines a large portion of the KSA output. Second, a portion of the

key input to the KSA was found to be “exposed to the attacker.” Fluhrer, Mantin, and Shamir found that

because the same secret portion of a key is used with different exposed values, an attacker could

reconstruct the secret portion by analyzing the first word of keystreams. In short, correlations were

found to exist between the WEP key and the “random” keystream and these similarities were found to
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 18
04/18/08

be revealed to attackers (Fluhrer, S., Mantin, I., & Shamir, 2001).

By collecting encrypted packets with certain IVs (called “weak” IVs), an attacker could backtrack

through the encryption process and crack the key (Hendrickson & Piotrowski, n.d., p. 4). By taking

advantage of these leaked correlations, one could derive the WEP key by capturing enough encrypted

traffic. Fluhrer, Mantin, and Shamir even formulated an attack that they proposed may crack the

encryption scheme. A month later, Adam Stubblefield successfully implement the attack (AirSnort,

2004). Stubblefield stated that from 5,000,000 to 6,000,000 encrypted packets were needed to crack a

104-bit WEP key (Stubblefield, Ioannidis, & Rubin, 2001). The cracking utilities AirSnort and

WEPCrack were soon released on the Internet, and became the first publicly available implementations

of the FMS attack (AirSnort, 2004).

The findings of Fluhrer, Mantin, and Shamir changed everything. It was apparent now that WEP was

fundamentally flawed (Gast, 2002). The attacks on WEP didn't stop, however. As time went on, new

and better ways to crack WEP were discovered. In 2004, a hacker known as KoReK substantially

improved upon the FMS attack, reducing the necessary number of captured packets to crack a 104-bit

WEP key down to 500,000 to 2,000,000 packets. A year later, Andreas Klein demonstrated that there

are even more correlations between the RC4 keystream and the key than Fluhrer, Mantin, and Shamir

had found. Based on Klein's work, German researchers Andrei Pychkine, Erik Tews, and Ralf-Philipp

Weinmann formed a new attack referred to as aircrack-PTW in April of 2007. Using this optimized

attack, one can crack a 104-bit WEP key with 40,000 to 85,000 packets at a success rate of 50 to 95

percent, respectively. The authors of the attack claim that it takes a mere 3 seconds on a Pentium-M 1.7

GHz computer system to crack the key from captured packets (Tews, Pychkine, & Weinmann, 2007a).
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 19
04/18/08

The name of the team's paper revealing the new attack is “Breaking 104 bit WEP in less than 60

seconds,” (Tews, Pyshkin, & Weinmann, 2007b) Clearly, WEP is woefully inadequate and very broken.

Even to this day, new attacks are being formulated against WEP with great success. One example is the

“Latté Attack” or “AP-less WEP cracking” found in October of 2007. In this attack, an attacker can

take advantage of the Windows Wireless Stack and WEP flaws on a client system to obtain the WEP

key from a corporate network (Phifer, 2007e).

The long and the short of these vulnerabilities is that WEP has broken encryption, flawed

authentication methods, and a flawed message integrity system, making it vulnerable to a wide range of

wireless attacks. It is ironic that the three main aims of the IEEE 802.11 group, confidentiality, access

control, and data integrity, are all broken in WEP. In summary, WEP has many vulnerabilities,

including:

● Brute force attacks (with a 40-bit key, such an attack is feasible)

● Dictionary attacks (the key can be guessed if weak)

● Keystream attacks (keystream cracking allows eavesdropping & message injections)

● Statistical correlation attacks (FSM, KoReK, and PTW attacks allow key cracking)

● Man-in-the-middle attacks (due to one-way authentication)

● Denial of service attacks (through disassociation frames)

● Weak message authentication (the ICV can be modified, packet injections possible)

● Flawed shared-key authentication (keystream exposed)

● Poor key management (the use of a single key by many users)

Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 20
04/18/08

D. MAC FILTERING AND CLOSED NETWORKS HAVE THEIR OWN ISSUES

While MAC filtering is a viable choice for many small businesses and home environments, its

usefulness disappears in larger wireless deployments. It becomes overly cumbersome to keep static lists

of every wireless device in use at a large organization. In addition, this method of access control can be

easily faked, or “spoofed.” Simply looking at wireless packets provides attackers with valid MAC

addresses. By typing two root commands on a LINUX system, an attacker can easily spoof a MAC

address and can circumvent this access control method. On a Windows system, the registry can be

modified to change the MAC address (Shimonski, 2003).

Lucent's proprietary “closed network” concept limits access to only those who know the SSID by

disabling SSID beaconing on the access point. While this sounds like a good idea, it is well known that

SSIDs are broadcast without encryption in many management messages. With a sniffer, anyone can

determine the SSID and join a closed network. Thus, because the SSID is not truly a secret, closed

networks are not really closed (Arbaugh, Shankar, Wan, & Zhang, 2002).

Although it has long been recommended in security circles to disable SSID beaconing wherever

possible, today it is recommended to leave SSID beaconing enabled. With SSID beaconing disabled, all

clients are forced to constantly send out probe requests as they attempt to locate the hidden network.

These probe requests contain the SSID and are sent out from clients wherever they are - even at remote

locations. This is a potential security threat as an attacker could capture these probes, determine the

SSID, and then set up a rogue access point with the captured SSID (creating an “evil twin”). From here,
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 21
04/18/08

the client may attempt to connect to the fake network, and the attacker can capture the authentication

traffic for use in future attacks (Ou, 2008). Microsoft now “strongly” recommends leaving SSID

beaconing enabled so that clients will not automatically send out these probe requests (Microsoft

Technet, 2006). In addition, disabling SSID beaconing will increase the traffic overhead of the network

and can degrade performance (Phifer, 2003; Moskowitz, 2003, p. 3). In summary, closed networks are

more trouble than they are worth.

On the other hand, disabling SSID beaconing will stop many casual snoopers from discovering the

network. Disabling SSID beaconing may be desirable for small home-based wireless networks with

clients that are connected for short periods and do not leave the immediate area (Phifer, 2003). Many IT

experts argue, however, that the threat posed by probe requests broadcast by clients negates any benefit

gained by “hiding” the SSID (Ou, 2008).

E. WI-FI'S WPA TO THE RESCUE

After serious WEP vulnerabilities were discovered in 2001, the wireless world was looking for a

solution that would address WEP's shortcomings. WEP2 and Dynamic WEP are two of these attempted

fixes. WEP2 attempted to address WEP's problems in two ways: by increasing the WEP key to 128 bits

instead of 64 and by implementing a new authentication scheme known as Kerberos. Unfortunately,

these two steps did not substantially improve the security of WEP. Because WEP2 had its own share of

weaknesses it was rarely implemented (Ciampa, 2006, p. 293).

Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 22
04/18/08

Dynamic WEP was another proposed solution. Dynamic WEP frequently rotates keys, making packets

harder to crack. Keys can be set to rotate at a specified period, such as every 15 minutes. While

Dynamic WEP largely overcomes the weaknesses of WEP's weak IV's, it is not a total security solution

and is still susceptible to man-in-the-middle and denial-of-service attacks. Because of this, it was not

widely implemented either (Ciampa, 2006, pp.293-294).

In March of 2001, IEEE TGi split into two groups, one of which would create a new wireless security

scheme. It was three years until IEEE 802.11i, the new wireless security model, was finally ratified.

During this time, the Wi-Fi Alliance, a consortium of equipment manufacturers and software providers,

decided that wireless security could wait no longer and developed Wi-Fi Protected Access (WPA).

WPA was introduced in October of 2003 with the goal of providing security for present and future

wireless devices. WPA is a subset of the 802.11i standard and implements security through encryption

and authentication (Ciampa, 2006, pp. 49, 295, & 297)

WPA was designed to replace WEP and was designed to fit within the WEP procedure with a minimal

amount of change (Ciampa, 2006, p. 305). WPA uses the Temporal Key Integrity Protocol (TKIP)

rather than a static encryption key. Instead of using a pre-shared key which creates a keystream, WPA

uses a pre-shared key to serve as the seed for generating the encryption keys. WPA increases the IV

space to 48 bits (now called the “sequence number”) and uses a long, 128-bit per-packet key to

dynamically encrypt each packet with a different key (Ciampa, 2006, pp. 304, 306, & 297). This solves

the IV collision problem and eliminates the correlation weaknesses of WEP.

WPA also eliminates the Cyclic Redundancy Check (CRC) function and replaces it with Message
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 23
04/18/08

Integrity Check (MIC). WEP's CRC system does not adequately protect the integrity of data, as one can

alter both the packet and the CRC value and craft a valid packet. WPA's MIC implements a strong

mathematical function computed by the transmitter and receiver. If the computed MIC value does not

match the packet's MIC value, the packet is discarded. This makes packet modification extremely

difficult, if not impossible (Ciampa, 2006, p. 298). Refer to Figure-4 for a description of the WPA

encryption process.

Step 1 Step 2

Sender's Sequence
Temporal Key XOR MAC = Value1 + Number
= Value2

PRNG = Keystream
Plaintext
XOR

Step 3 MIC = Plaintext MIC

MIC Key Ciphertext

Sender's MAC
Receiver's MAC
Figure-4. The WPA Encryption Process.
Source: (Ciampa, p. 305)

In step 1, a temporary key is generated and is XORed with the sender's MAC address, creating Value1.

In step 2, Value 1 is mixed with a sequence number (which replaces the IV) to produce Value2, the per-

packet key. Value2 is then run through a Pseudo Random Number Generator (PRNG), as in WEP. The

output of the PRNG is the keystream, which is used for encryption. In step 3, instead of inputting the

plaintext message into a CRC function, the MIC Key and the sender's and receiver's MAC addresses
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 24
04/18/08

are all entered into a MIC function. The result is the plaintext message with a MIC key appended.

Finally, the plaintext message and the MIC key are XORed with the keystream, producing the

ciphertext. The receiver uses its copy of the pre-shared key to backtrack through the encryption process

and produce the original plaintext message (Ciampa, 2006, p. 305).

TKIP in WPA is composed of three major components:

1) the MIC, which protects against packet forgeries

2) an IV sequence, which protects against replay attacks by sequencing each packet, and

3) TKIP key mixing, which uses temporary per-packet keys with limited lifespans

TKIP forms the heart of WPA. Compared with WEP, WPA requires similar computational effort and

superior security (Ciampa, 2006, p. 306). Because the encryption process is so similar to WEP,

implementing WPA can be as simple as upgrading client software and updating the firmware of older

access points (Ciampa, 2006, p. 299).

WPA offers two authentication modes: 1) pre-shared key (PSK) authentication based on a secret key

shared among devices, and 2) authentication via a dedicated authentication server. The pre-shared key

mode is intended for personal and small home /office environments (SOHO) whereas authentication

via an authentication server is designed for enterprise networks. The IEEE 802.1x standard defines how

authentication takes place with an authentication server, and is discussed in more detail in the next

section (Ciampa, 2006, pp. 304, 308).

Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 25
04/18/08

Although WPA provides comprehensive and robust security in a small home /office environment and

adequate security in an enterprise environment, it is not immune to attack. Brute-force dictionary

attacks are possible in WPA-PSK (Pre-Shared Key) networks. The popular hacker's tool “coWPAtty”

can be used to launch dictionary attacks against WPA-PSK networks (Wirelessdefence.org, 2006). To

avoid a successful dictionary attack, it is important to enter either a random sequence of at least 20

ASCII characters or a random sequence of at least 24 random hexadecimal digits as the passphrase (the

passphrase is used to create the shared secret) (Ciampa, 2006, p. 307). Words, which are not random,

should never be used in a passphrase.

WPA was meant to fit within the confines of the WEP structure. Because WPA was not created from

scratch, security vulnerabilities may be discovered in the future (Ciampa, 2006, p. 311). Today, it is

recommended to use WPA2 (discussed shortly) instead of WPA whenever possible.

F. IEEE 802.11I & WPA2 FINALLY ARRIVE

A year after WPA was introduced, the IEEE 802.11i standard was finally ratified. Also referred to as

Robust Security Network (RSN), IEEE 802.11i is considered to be a solid security model, providing a

high level of wireless security. Instead of a stream cipher such as RC4, IEEE 802.11i uses the block

cipher Advanced Encryption Standard (AES) for encryption. AES breaks the plaintext down into blocks

of 8 to 16 bytes and manipulates these blocks. Through a series of iterations, or rounds, the bytes are

rearranged and substituted. Multiplication is then performed on the changes. The result is a much

stronger encryption scheme, practically immune from modern brute-force attacks (Ciampa, 2006, p.
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 26
04/18/08

295).

Authentication and key management in IEEE 802.11i are provided by the IEEE 802.1x standard. IEEE

802.1x was originally designed for wired networks, and operates using port security. A particular

device requesting access to the network is restricted from doing so or receiving any traffic from that

network until its credentials are verified. IEEE 802.1x blocks all ports to a device until it is

authenticated by an authentication server. An authentication server is a server which verifies the

identity and privileges of requesting devices (Ciampa, 2006, p. 296). This could be a dedicated server

or an integrated module in the access point (Chen, Jiang, & Liu, 2005).

An example of an authentication server would be a RADIUS (Remote Authentication Dial-In User

Service) server. Refer to Figure-5 for a simple description of the authentication process in a RADIUS-

based IEEE 802.1x system.

Step 1: Supplicant asks to join

Step 4: Authentication
Step 2: Authenticator asks for credentials server verifies
credentials

Step 3: Supplicant sends credentials, forwarded to Authentication server

Step 5: Supplicant is
approved or denied

Supplicant Authenticator
Authentication Server
(Access Point)
(RADIUS server)

Figure-5. IEEE 802.1x authentication

Source: (Ciampa, p. 297)
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 27
04/18/08

First, the requesting device, the supplicant, associates with the access point, often called the

authenticator. The authenticator acts as a proxy for the user's authentication, and completely blocks off

the protected network from the supplicant. In fact, it only opens a single port to the supplicant with

itself. The supplicant begins by asking for network access. The authenticator asks for its credentials,

and receives them from the supplicant. The authenticator then forwards on the request and credentials

to the authentication server, which in this case is the RADIUS server. The authentication server first

evaluates if the access point is allowed to make such requests. If it is, the RADIUS server looks up the

user's credentials and determines if the user is in fact permitted the access that is being requested.

Depending on the authentication method and the credentials used, challenge text may be sent back to

the supplicant, which the supplicant must answer with the correct value. Once the authentication server

has decided whether or not the supplicant should be granted access, it sends its answer to the

authenticator, which is relayed to the supplicant (Ciampa, 2006, pp. 296-297 & 309-310; Fleishman,

2003). Using this method, the authentication server is never directly contacted. This affords a greater

degree of security for the authentication server, which contains the valuable database of user

credentials.

IEEE 802.1x uses the Extensible Authentication Protocol (EAP) to handle the presentation of user

credentials, which may be digital certificates, usernames and passwords, smartcards, or secure IDs.

Many EAP types are available for implementation. Common EAP types that are used include EAP-

Transport Layer Security (EAP-TLS), EAP-Tunneled Transport Layer Security (EAP-TTLS), and

Protected Extensible Authentication Protocol (PEAP) (Chen, Jiang, & Liu, 2005, p. 28; Wi-Fi Alliance,

2003, p. 5).
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 28
04/18/08

Using an authentication server makes user administration easier and can thwart man-in-the-middle

attacks due to two-way authentication. By using an EAP implementation such as EAP-TLS, EAP-

TTLS, or PEAP, for example, both the client and server will mutually authenticate each other and man-

in-the-middle attacks become much harder to wage (Intel Corporation, 2007).

In September of 2004, the Wi-Fi Alliance introduced Wi-Fi Protected Access 2 (WPA2), based on the

IEEE 802.11i standard. WPA2 implements AES encryption and provides an even stronger security

system for WLANs than WPA. WPA2 is similar to 802.11i, but it allows for greater interoperability

with WPA (Ciampa, 2006, p. 299). Encryption under WPA2 uses the AES-CCMP (Counter Mode CBC-

MAC Protocol) protocol. The AES algorithm deals with blocks of 128 bits, but the length of the cipher

keys and rounds of encryption can be changed. In the IEEE 802.11i/WPA2 implementation of AES, a

128-bit key is used, with each round iterated 10 times (Ciampa, 2006, p. 307).

Like WPA, WPA2 offers two security modes: 1) pre-shared key authentication based on a shared secret,

and 2) authentication by an authentication server (IEEE 802.1x). Pre-shared key authentication is

intended for personal and small office (SOHO) use where an authentication server is unavailable.

(Ciampa, 2006, pp. 306, 311). Like WPA, WPA2 networks using a pre-shared key are vulnerable to

dictionary attacks (Phifer, 2007b). It is important to make the secret passphrase as long and as random

as possible (at least 20 characters long), with a healthy mix of various random characters (Ciampa,

2006, pp. 306,307, & 311).

The authentication method recommended for enterprise use is IEEE 802.1x - currently the strongest

authentication method available. IEEE 802.1x is currently very attractive to enterprise environments in
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 29
04/18/08

which many distinct wireless users must access the network. IEEE 802.1x allows for centralized

wireless access management (Ciampa, 2006, pp. 306 & 311).

In summary, wireless security has been a bumpy road. First there was WEP, but it had serious

vulnerabilities. This was followed by a more secure WPA, and finally, a very secure WPA2. With a

basic understanding of these security systems, we can now focus on how each of these systems holds

up against common wireless attacks.

Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 30
04/18/08

IV.COMMON WIRELESS ATTACKS

The following section lists some common wireless attacks and how network security solutions handle

these attacks. Best-selling author Robert Shimonski (2003) classifies wireless attacks into four main

categories:

1. Passive Attacks (eavesdropping packets)

a. Brute force attacks

b. Brute force dictionary attacks

c. Statistical correlation attacks

2. Active Attacks

a. Unauthorized Access (unwanted file sharing, adding/changing/deleting data)

b. Spoofing

c. Denial of Service attacks

3. Man-In-The-Middle attacks

4. Jamming attacks (also Denial of Service)

Passive attacks are performed by passively capturing wireless traffic. Thus, passive attacks use

eavesdropping to further their purpose. If the network is already open, eavesdropping could be

considered an attack in and of itself. If valuable information is transmitted over an open wireless

connection, for example, an attacker's eavesdropping would have already achieved the purpose of

capturing that information (Shimonski, 2003).

Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 31
04/18/08

The freeware program “Netstumbler” is a popular wireless program that is commonly used to locate

wireless networks. It can identify the Service Set Identifier (SSID), determine the encryption used, and

even determine the manufacturer of the access point. If the SSID of the access point is hidden, other

programs such as “Ethereal,” WildPacket’s “AiroPeek,” or “Kismet” can be used to locate a WLAN

and determine the SSID (Kismet, n.d.; Shimonski, 2003).

Often, the capturing of wireless network traffic is but the first step in an attack. Brute force attacks are

attacks which attempt to break the encryption of captured traffic through brute force, trying every

possible key combination. At lower bit strengths, such as WEP's original 56-bit encryption, such

attacks become feasible with modern equipment. Higher bit strengths, such as 128-bit encryption, make

these attacks less feasible.

A particularly popular kind of brute-force attack is the dictionary attack, also called the “offline”

dictionary attack. It differs from the raw brute force method in that it uses a dictionary of commonly

used words to attempt to break the encryption key. Such attacks can be very effective if a weak

passphrase is used. For example, if the secret passphrase is “secret,” a dictionary attack would attempt

different commonly used words in encryption and compare the result with the captured traffic. When

there is a match (in this case, “secret”), the passphrase (key) is cracked. WPA/WPA2 networks using a

pre-shared key are vulnerable to this attack. As a matter of fact, the dictionary attack is the only known

cryptographic vulnerability of WPA/WPA2-PSK networks (Aircrack-ng, 2008; Shimonski, 2003).

Statistical correlation attacks exploit flaws in the encryption scheme to crack the key from captured

traffic. The most widely known correlation attacks are the WEP attacks FMS, KoReK, and PTW. Flaws
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 32
04/18/08

in WEP reveal correlations between the key and the keystream, and from these correlations, attackers

can eventually crack the key. WPA and WPA2 networks are currently believed to be immune from

these attacks (Aircrack-ng, 2008).

Active attacks are attacks which not only receive wireless traffic but also transmit wireless traffic,

taking an active role in the targeted wireless network. After gleaning information from a passive attack,

an active attack can be waged. Unauthorized access, spoofing, and denial of service attacks can be

considered to be active attacks (Shimonski, 2003).

Unauthorized access is the use of resources by unauthorized parties. If a wireless network is open, with

no encryption, attackers may be able to access file shares, add or delete files, and introduce malware

onto connected computer systems. The TJX break-in in our introduction is an example of the

unauthorized use of data. WEP has known flaws that enable attackers to crack the key and gain entry

into the wireless network. WEP's shared-key authentication system is also flawed, and through

eavesdropping, an attacker can observe the authentication sequence and spoof an authenticated

connection (Shimonski, 2003).

Because the goal of many wireless attacks is the unauthorized access of resources, sound security

practices are needed to protect resources from wireless attack. This includes setting the proper file and

folder permissions on computers connected to the wireless network as well as implementing sound

wireless security, such as WPA/WPA2. In addition, a firewall between the access point and the wired

network will help to protect the internal network from possible intrusions. Installing software firewalls

on wireless clients helps to protect them from unauthorized wireless access.

Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 33
04/18/08

Spoofing is the impersonation of an authorized station to gain unauthorized access to a wireless

network. This is commonly done to circumvent MAC filtering. By changing a registry entry on a

Windows system or by entering two root commands on a LINUX system, an attacker's MAC address

can be spoofed to match that of a valid user. This is a known weakness of MAC filtering (Shimonski,

2003).

Active denial-of-service attacks attempt to deny service to authorized parties. Over a wireless network,

these attacks can be identical to their wired counterparts and include ping floods, SYN attacks,

fragment attacks, and Distrbuted DoS (DDoS) attacks. Open networks are vulnerable to these denial-

of-service attacks, and because WEP is easily cracked, it too is vulnerable (Shimonski, 2003).

Fortunately, due to strong encryption and authentication,WPA/WPA2 based networks are protected

from these classic attacks. However, WPA networks can be vulnerable to a sly version of the DoS

attack. WPA access points have the option of stopping all transmissions (disassociating all devices and

preventing new associations) for one minute if two packets with malformed MICs are detected within

one minute. Therefore, if this option is enabled, an attacker could craft malformed packets with bad

MICs for the sole purpose of shutting down the network; a denial of service attack. Since this MIC

countermeasure is optional, it can be disabled, preventing this particular DoS attack. (Ciampa, 2006, p.

298; Maufer, 2003, p. 281).

One wireless form of a denial of service attack is the kicking off of legitimate users by sending false

disassociation frames. When the access point receives these frames, it will disconnect the (supposed)
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 34
04/18/08

sender from the network. WEP and WPA/WPA2 networks are vulnerable to these attacks, and they can

be very difficult to identify and defend against. Special RF scanners can be used to detect and pinpoint

unauthorized wireless devices that may send out these frames (Ciampa, 2006, p. 383).

Man-in-the-middle attacks are a class of attacks that set up rogue (illegitimate) access points within

range of wireless clients for the purpose of acting as a “middle man” for clients. When clients see the

rogue access point (also called an “evil twin”), the SSID matches the legitimate access point and they

mistakenly join it instead of the true access point. From here, the attacker can forward on their requests

to the true access point, but he can monitor or change their traffic. It has been common in the past for

attackers to use two network cards in their laptops, one serving as the phony access point, and the other

forwarding on traffic to the real access point. Attackers may hide their rogue access points in closets or

under desks near the true access point (Shimonski, 2003).

Historically speaking, man-in-the-middle attacks have always been a problem with wireless networks.

WEP-based networks are very vulnerable to man-in-the-middle attacks. WPA and WPA2-based PSK

networks are not immune to man-in-the-middle attacks because of their lack of mutual authentication

(Phifer, 2007a). WPA and WPA2-based networks using IEEE 802.1x enterprise authentication are

protected from man-in-the middle attacks if an EAP type supporting mutual authentication is used

(Intel Corporation, 2007).

Rogue access points themselves are a major security threat for wireless networks, even if they are not

set up by attackers for a man-in-the-middle attack. Rogue (unauthorized) access points can be set up by

employees to make networking easier. Because unauthorized access points connected to the LAN open
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 35
04/18/08

a new door into the internal network, they effectively bypass established security mechanisms and

present a clear danger. WPA/WPA2 clients are not inherently protected from joining unauthorized

access points, but employing IEEE 802.1x authentication can prevent users from mistakenly joining

phony access points (Bulk, 2006; Phifer, 2007a).

To protect against rogue access points, employees should be banned from setting up their own access

points and the organization's airwaves should be periodically scanned for unknown access points

(Cisco Systems, Inc., 2007). A Wireless Intrusion Prevention System, which continuously scans for

such activity, can be very useful in detecting unauthorized access points (Phifer, 2007a).

Finally, jamming attacks use RF transmissions to flood the airwaves with noise and block wireless

network transmissions. This is an inherent problem in wireless networks because traffic collisions are

avoided by transmitting devices. If wireless traffic (or a traffic-like transmission) is detected, stations

must wait until the airwaves are clear again before they can transmit. Cordless phones and microwave

ovens are known to interfere with wireless LANs. In addition, Bluetooth devices and other wireless

hardware can cause wireless jamming. Jamming may be intentional or unintentional and is relatively

rare. To prevent jamming, policies must be made banning the use of interfering equipment near the

wireless network (Ciampa, 2006, p. 280; Shimonski, 2003).

Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 36
04/18/08

V. WIRELESS NETWORKS USED TODAY

With an understanding of the different wireless security mechanisms available and the common

wireless attacks waged on wireless networks, we can now turn our attention to examining the wireless

networks used today. There are many variants of wireless networking security in use today. The most

prevalent systems found in operational use are:

1. No Security

2. WEP

3. WPA/WPA2-PSK

4. WPA/WPA2 with RADIUS

5. VPN (IPSEC, SSH, or TLS / SSL over WLAN)

Wireless LANs with no security are the most vulnerable to attack. They are in effect “wide open.”

There is no authentication of users, no encryption of transmissions, and no security measures in place.

This is a wireless attacker's dream. With no security measures, an attacker is free to eavesdrop, use the

Internet connection (if available), access any file shares, or perform other malicious actions.

There are many reasons why wireless networks are left unsecured - one is ignorance. Because wireless

networks are so easy to install, many people just plug them in and use them. While the networks do

“work,” they also form a soft spot that is just waiting to be taken advantage of.
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 37
04/18/08

As we have seen, WEP – secured wireless LANs are not truly secure. While WEP's vulnerabilities have

been known for years, a large percentage of current companies still use WEP (Phifer, 2007d). For

many, the reason comes down to cost. More efficient security mechanisms such as WPA must be

implemented within the device itself. While the firmware on most WEP-based access points and

wireless network interface cards can be upgraded to support WPA, some devices cannot be upgraded to

support WPA. If the device does not support WPA, the administrator must replace the device to

implement WPA (Wi-Fi Alliance, 2003, p. 2).

Many early point-of-sale retail computer systems and scanners only support WEP (Fleishman, 2007).

The cost of replacing wireless network devices is a heavy burden to many small businesses who cannot

invest in more expensive equipment. Other organizations are simply ignorant of the security

vulnerabilities of WEP. It is also easier to setup a wireless network the way it comes; without upgrades.

For these reasons, a number of organizations still use WEP in their wireless networks. While WEP is

insecure, it is definitely better than no security at all.

Before WPA became available, the conventional wisdom for securing wireless networks was to treat

them as insecure, placing access points outside of a firewall and then tunneling VPN connections

through to the wired network. Any other traffic would be insecure (Fleishman, 2003). While it certainly

is more secure to treat a wireless connection as untrusted, and organizational firewalls are extremely

important, WPA has allowed many organizations to treat wireless networks with increased trust. If

properly configured, WPA – based wireless LANs are relatively safe from wireless attacks. The

symmetric keys used in WPA-PSK cannot be easily cracked if a strong passphrase (which is used to

create the pre-shared key) is used. Attackers are currently limited to dictionary attacks in WPA. The
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 38
04/18/08

security of the symmetric key depends on the strength of the passphrase used; if it is easy to guess, the

key may be easily cracked. This is why a strong passphrase is so important.

WPA – based wireless LANs using IEEE 802.1x authentication are also relatively secure from wireless

attacks. It is important to remember, however, that WPA was not built from the ground up and does not

offer the highest security protection possible.

WPA2 – based WLANs are highly secure networks, fit for enterprise-level adoption. The AES-CCMP

encryption standard ensures that encrypted packets cannot be cracked. Just as in WPA, however, in

personal security / pre-shared key mode, a symmetric key could also be cracked if it is easy to guess.

WPA2 is the most secure wireless network configuration in use today.

In enterprise environments, WPA/WPA2 networks utilize IEEE 802.1x authentication and use an

authentication server, such as a RADIUS server, to authenticate clients. This allows administrators to

place all wireless users into a database and configure their access rights from a central location. Central

management makes security easier to implement, makes network usage easier to track, and increases

security because a network-wide policy can be applied at a central server (Ciampa, 2006, p. 310).

VPNs over WLANs are not as common as the other types of wireless security mentioned, but they are

used by many organizations for increased security over a WLAN (Intel Corporation, 2007). A secure

tunnel allows for private communications over an insecure medium. As an example, one could use SSH

server and the SSH client “PUTTY” to establish a secure tunnel over a wireless network. Traffic could

then be securely sent through this encrypted tunnel. One must be cautious in implementing a VPN over
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 39
04/18/08

a wireless link, however. VPNs do not solve all WLAN security problems. User error in a VPN

application could have catastrophic consequences, and could result in sensitive data not being enclosed

in the tunnel. In addition, the cost of the equipment and continued management can be high and the

complexity of the network increases, making things harder for the administrator (Intel Corporation,

2007).
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 40
04/18/08

VI.ROADS TO SECURITY – COMMONSENSE GUIDELINES

Depending on the current wireless network, security needs, and the available budget, there are different

roads to improving wireless security. An organization must decide, based on its security policy, how to

implement an effective balance between operating needs and security. The following recommendations

are for organizations who want to secure their WLANs which are connected to a private network (basic

and extended service sets). Wireless hotspots and ad-hoc networks require different security and

operational needs.

1. Basic Security Measures

2. Interim Security Measures (for little to no resources)

3. WPA/WPA2-PSK

4. WPA/WPA2-Enterprise

5. VPN, SSL over WLAN

6. High Security WLAN Precautions

Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 41
04/18/08

Basic Security Measures

First among the steps to securing a wireless network are basic security measures, which should be

implemented regardless of the security system in place.

1. First scan the area for conflicting access points. This will prevent Denial of Service issues down

the road (Dowler, 2007, p. 9).

2. Place the access point in a favorable location. The access point should be placed where the

transmissions will be broadcast in the desired location and range (Cox, 2003). By reducing the

power of the access point or by using a directional antenna, one can reduce the range of the

wireless network.

3. Change the default SSID of the access point to an abstract and random name. Naming a

wireless network after a business or a family name is a bad idea. Attackers could use the SSID

to determine who the network belongs to, and with more effort, could determine possible

weaknesses (Dowler, 2007, p. 10).

4. Change the default username/password of the access point. Follow strong password creation

procedures by using various characters and making the password as long as possible (Dowler,

2007, p. 9).

5. Leave SSID beaconing enabled, as this will prevent clients from constantly sending out probe

requests with the SSID (Microsoft Technet, 2006). On clients, be sure to configure the network

as a broadcast network so that probe requests will not be sent out (Rubens, 2007). If you do

decide to disable SSID beaconing, set the access point to respond only to probe requests with

the SSID, if possible (Macaulay, 2002, p. 11). Disabling SSID beaconing may not be possible
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 42
04/18/08

on WLANs with roaming users (Ciampa, 2006, p. 302).

6. Enable MAC filtering, if possible. This may not be feasible in large deployments or possible in

roaming networks ( Macaulay, 2002, p. 8; Phifer, 2007c).

7. Disable unused network shares on clients and disable peer-to-peer wireless connections on

clients if an access point is used. Ensure that only trusted connections are listed on the client's

wireless connection profile (AirTight Networks, Inc., 2007).

8. Ensure clients are properly protected with software firewalls.

9. Update the drivers of wireless network interface cards (NICs) in clients and update the firmware

of access points. Keep the drivers and firmware up to date.

10. Regularly check device logs for unusual activity.

11. Periodically perform site surveys, or “sweeps” for rogue access points or other banned wireless

devices (Cisco Systems, Inc., 2007).

12. Turn off access points when the wireless network is not in use.
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 43
04/18/08

Interim Security Measures

If WEP is the only option, interim security measures can be used until the network is upgraded. Interim

security measures, also called a “transitional” security system by author Mark Ciampa (2006), operates

on the idea that it is best to implement as many security measures as possible. Even if one by one these

measures could be evaded, together they make life more difficult for attackers and will turn away most

undetermined snoopers. At any rate, it is important to consider these steps as temporary until a more

secure wireless system can be installed. Even after taking these security measures for WEP, WEP is still

open to attack. It is imperative to replace WEP in your organization. Until WEP can be replaced:

1. Implement Basic Security Measures (discussed previously).

2. Enable a minimum of 104-bit WEP encryption, using a key made with random hexadecimal

characters. Avoid passphrase generators, as the keys they produce may not be truly random

(Ciampa, 2006, p. 303; Phifer, 2007c).

3. Rotate shared WEP keys at frequent intervals, such as quarterly (Phifer, 2007c).

4. Disable shared-key authentication; use open authentication. Shared key authentication is so

flawed that Microsoft now recommends disabling it, opting instead for open authentication with

WEP encryption (Microsoft Technet, 2006).

5. Treat the wireless network segment as insecure and place a firewall between the access point

and the protected network (Macaulay, 2002, p. 8). Any data sent over the wireless network with

only WEP encryption for protection is not to be trusted.

6. Seriously consider implementing a VPN; see VPN Security Measures.

Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 44
04/18/08

WPA/WPA2-PSK Measures

Implementing WPA instead of WEP brings a substantial boost in security. If wireless devices do not

support WPA2, WPA is the best wireless security option available. WPA/WPA2 networks can use a

shared secret (a secret passphrase) in pre-shared key mode (PSK). This mode is intended for small

home/office (SOHO) use of 10 devices or less (Ciampa, 2006, p. 304).

1. Implement Basic Security Measures (discussed previously)

2. Use a strong passphrase. Either use a random sequence of at least 20 ASCII characters or a

random sequence of at least 24 hexadecimal digits (Ciampa, 2006, p. 307). Make the passphrase

as long and random as possible. If using ASCII characters, use numbers, letters with case

changes, and other ASCII characters. WPA/WPA2 are vulnerable to dictionary attacks (Phifer,

2007b). The longer and more random the passphrase, the better.

3. Ensure the SSID is a unique, uncommon name, different than the default. The SSID is mixed

with the shared secret to create the key. Hackers use tables of common SSIDs to speed up the

cracking process (Phifer, 2007b).

Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 45
04/18/08

WPA/WPA2-Enterprise Measures

1. Implement Basic Security Measures (discussed previously).

2. Use a robust EAP implementation that supports mutual authentication, such as EAP-TTLS,

EAP-FAST, or PEAP (Intel Corporation, 2007).

3. Disable unused EAP types on the RADIUS server (Wright, 2007).

4. Ensure that clients properly identify and authenticate the authentication server. The default

operating system settings may accept all root certificate authorities. Clients should only accept

digital certificates from the approved authentication server. These settings can be set through

group policy objects in Windows (Wright, 2007).

5. Avoid using EAP types that are vulnerable to brute force dictionary attacks, such as EAP MD5,

EAP GSS with Kerberos V, or Cisco LEAP (Aboba, 2006).

Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 46
04/18/08

VPN Security Measures

1. Implement Basic Security Measures (discussed previously).

2. Ensure that the VPN software is being correctly implemented.

3. Educate users how to properly use the tunneling software.

4. Place a firewall between the access point and the protected network, allowing only traffic over

tunneled ports.

5. Place software firewalls on wireless clients that block traffic not traveling over tunneled ports.

6. Keep tunneling applications up to date.

7. Routinely check the VPN logs for attempted break-ins or unsuccessful logins.

8. Consider implementing a NIDS (Network-based Intrusion Detection System) behind the

firewall. A network-based IDS can detect attack patterns or unusual traffic patterns and sound

an alert (Whitman & Mattord, 2004, p. 289).

9. Periodically perform checks on the VPN system.

Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 47
04/18/08

High Security WLAN Security Precautions

In addition to using a WPA/WPA2 secured network, the following measures can be implemented in

networks that require increased security. Enterprises handling sensitive information and organizations

especially concerned about security may require additional security measures.

1. Treat the wireless network as being insecure. Traffic flowing over the wireless medium is

potentially vulnerable to attack, and the wireless medium may serve as an entry point for

attacks.

2. Place a firewall between the access point and the protected network and only allow required

traffic to pass.

3. Place an IDS behind the firewall to monitor for wireless intrusions. Wireless break-ins may

cause anomalous traffic patterns. Installing a network-based IDS will detect these patterns and

sound an alert. Implementing an IDS can take considerable effort, however, so be warned

(Whitman & Mattord, 2004, p. 289).

4. Consider implementing a WIDS (Wireless Intrusion Detection System) / WIPS (Wireless

Intrusion Prevention System). Such a system is composed of wireless sensors which scan the

airwaves for rogue access points, known attack patterns, or odd wireless behavior (Ciampa,

2006, p. 315). A WIDS / WIPS continually monitors the security of a wireless network and can

take appropriate actions under specific conditions (DigitalAir Wireless Networks, 2006).

5. Consider installing network access control software on clients. A system such as

SpectraGuard® SAFE from AirTight Networks will allow clients to create a profile for trusted
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 48
04/18/08

zones and will stop clients from accidentally connecting to foreign wireless devices (AirTight

Networks, Inc., 2007).

6. Frequently check device logs for unusual activity.

7. Regularly perform site surveys, or “sweeps” for rogue access points or other banned wireless

devices.

8. For wireless networks transmitting especially sensitive data, implement a VPN on top of the

WPA/WPA2 architecture; see VPN Security Measures.

Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 49
04/18/08

VII.THE BATTLE CONTINUES

While security for wireless networks has come a long way since the original IEEE 802.11 standard was

adopted in 1997, it is important to remember that wireless security is an ongoing concern. One of the

most recently discovered wireless vulnerabilities has been the disclosure that certain wireless NIC

drivers can be compromised by bombarding the wireless NIC with floods of wireless packets. This

attack technique, called “fuzzing,” floods the wireless card with so much traffic that it can cause

programs to crash or even allow malicious software to execute. It also doesn't require the wireless NICs

to be connected to a network; the wireless NIC need only be turned on and functioning. Although

fuzzing is a technically tedious attack, with the proliferation of new attack scripts it is now an emerging

threat. This attack was demonstrated during Black Hat USA 2006, and is yet another reason to keep

wireless NIC drivers up to date (McMillan, 2006).

In the wireless arena, the battle for security is ongoing. The truth is, as long as a wireless network

broadcasts traffic on publicly available frequencies, it will be attacked. TJX was attacked and its

defenses failed because they were woefully inadequate. To keep such attacks from becoming

successful, it is important that the wireless administrator lock down the wireless network by following

the steps detailed in this paper. By tightening down wireless networks, would-be-attackers can be put

out of business, and your organization can be kept safe.

Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 50
04/18/08

VIII.WORKS CITED

Aboba, B. (2006, June 13). The Unofficial 802.11 Security Web Page.

Retrieved April 5, 2008, from

http://www.drizzle.com/~aboba/IEEE/

Aircrack-ng. (2008, March 27). faq [Aircrack-ng]

Retrieved April 4, 2008, from

http://www.aircrack-ng.org/doku.php?id=faq

AirDefense. (2007, November 15). AirDefense's Comprehensive Survey of 3,000 Retail Stores Finds

Many Wireless Data Security Vulnerabilities as Holiday Shopping Season Nears.

Retrieved April 4, 2008, from

http://www.airdefense.net/newsandpress/11_15_07.php

AirSnort. (2004, December 31). AirSnort Homepage: Introduction Section.

Retrieved April 4, 2008, from

http://airsnort.shmoo.com/
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 51
04/18/08

AirTight Networks, Inc. (2007, October 21). Caffe Latte Vulnerability Discovered by AirTight:

Underscores Urgent Need for Wireless Road Warriors to Adopt Best Practices

Retrieved April 5, 2008, from

http://www.airtightnetworks.net/news/pressrelease/news_company100.html

*Arbaugh, W.A., Shankar, N., Wan, Y.C.J., & Zhang, K. "Your 80211 wireless network has no clothes,"

Wireless Communications, IEEE [see also IEEE Personal Communications] , vol.9, no.6, pp.

44-51, Dec. 2002

Retrieved April 4, 2008, from

http://ieeexplore.ieee.org/iel5/7742/26001/01160080.pdf?isnumber=26001∏=STD&arnumber=

1160080&arnumber=1160080&arSt=+44&ared=+51&arAuthor=Arbaugh%2C+W.A.%3B+Sha

nkar%2C+N.%3B+Wan%2C+Y.C.J.%3B+Kan+Zhang

Ars Technica. (2002, July 17). Wireless Security Blackpaper.

Retrieved April 4, 2008, from

http://arstechnica.com/articles/paedia/security.ars/3

Borisov, N., Goldberg, I., & Wagner, D. Intercepting mobile communications: The insecurity of 802.11.

In Proceedings of MOBICOM 2001, 2001.

Retrieved April 4, 2008, from

http://citeseer.ist.psu.edu/borisov01intercepting.html
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 52
04/18/08

*Brown, B., "802.11: the security differences between b and i," Potentials, IEEE , vol.22, no.4, pp. 23-

27, Oct.-Nov. 2003

Retrieved April 4, 2008, from

http://ieeexplore.ieee.org/iel5/45/27781/01238689.pdf?isnumber=27781∏=JNL&arnumber=12

38689&arnumber=1238689&arSt=+23&ared=+27&arAuthor=Brown%2C+B.

Bulk, F. ( 2006, January 27). Crash Course: The ABCs of WPA2 Security. Network Computing.

Retrieved April 5, 2008, from

http://www.networkcomputing.com/showArticle.jhtml?articleId=177103376

*Chen, J-C., Jiang, M-C., & Liu, Y-w. "Wireless LAN security and IEEE 802.11i," Wireless

Communications, IEEE [see also IEEE Personal Communications] , vol.12, no.1, pp. 27-36,

Feb. 2005

Retrieved April 4, 2008, from

http://ieeexplore.ieee.org/iel5/7742/30466/01404570.pdf?isnumber=30466∏=JNL&arnumber=

1404570&arnumber=1404570&arSt=+27&ared=+36&arAuthor=Jyh-Cheng+Chen%3B+Ming-

Chia+Jiang%3B+Yi-wen+Liu

Ciampa, M. (2006). CWNA Guide to Wireless LANs (Second Edition).

Boston, Massachusetts: Thompson Course Technology.

Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 53
04/18/08

Cisco Systems, Inc. (2007). Addressing Wireless Threats with Integrated Wireless IDS and IPS.

Retrieved April 5, 2008, from

http://www.cisco.com/en/US/prod/collateral/wireless/ps5678/ps6521/prod_white_paper0900aec

d804f155b_ns386_Networking_Solutions_White_Paper.html

Cox, J. (2003, December 1). Build a secure enterprise WLAN. Network World.

Retrieved April 5, 2008, from

http://www.techworld.com/mobility/features/index.cfm?featureid=221

DigitalAir Wireless Networks. (2006). Wireless Intrusion Prevention Systems.

Retrieved April 5, 2008, from

http://www.digitalairwireless.com/wireless_intrusion_prevention_system.asp

Dowler, M. (2007, July 30). Beginners Guides: Securing A Wireless Network – PCSTATS.com.

In Beginners Guides.

Retrieved April 5, 2008, from

http://www.pcstats.com/articleview.cfm?articleid=1489&page=9

Fleishman, G. (2003, January 10). Key to Wi-Fi security. InfoWorld

Retrieved April 5, 2008, from

http://www.infoworld.com/articles/ne/xml/03/01/13/030113newifisec.html
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 54
04/18/08

Fleishman, G. (2007, October 17). New WEP Attack: Caffe Latte Hits Client, Not Access Point.

Wi-Fi Networking News.

Retrieved April 5, 2008, from

http://wifinetnews.com/archives/007993.html

Fluhrer, S., Mantin, I., & Shamir, A. "Weaknesses in the Key Scheduling Algorithm of RC4."

Presented to the Eighth Annual Workshop on Selected Areas in Cryptography, August 2001.

Retrieved April 4, 2008, from

http://www.crypto.com/papers/others/rc4_ksaproc.ps

Gast, M. (2002, April 19). Wireless LAN Security: A Short History. O'Reilly Media, Inc.

Retrieved April 4, 2008, from

http://www.oreillynet.com/pub/a/wireless/2002/04/19/security.html

Hendrickson, L. & Piotrowski, V. (n.d.). Wireless Security: from WEP to 802.11i.

Retrieved April 4, 2008, from

http://www.micsymposium.org/mics_2004/Hendrick.pdf

Intel Corporation. (2007, October 25). Wireless security - 802.1x and EAP types.

Retrieved April 4, 2008, from

http://www.intel.com/support/wireless/wlan/sb/cs-008413.htm
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 55
04/18/08

Kismet. (n.d.) Kismet.

Retrieved April 6, 2008, from

http://www.kismetwireless.net/

Macaulay, T. (2002, February 18). Hardening IEEE 802.11 wireless networks.

Retrieved April 5, 2008, from

http://www.mosteiro.hpg.ig.com.br/Material/Wireless/Others/Hardening_802.11.pdf

Maufer, T. A. (2003, October 17). A Field Guide to Wireless LANs for Administrators and Power

Users. Prentice Hall PTR.

McMillan, R. (2006, June 21). Researchers hack Wi-Fi driver to breach laptop. InfoWorld.

Retrieved April 5, 2008, from

http://www.infoworld.com/article/06/06/21/79536_HNwifibreach_1.html

Microsoft Technet. (2006, December 6). Recommendations for Small Office or Home Office Wireless

Networks.

Retrieved April 4, 2008, from

http://technet.microsoft.com/en-us/library/bb727047.aspx

Moskowitz, R. (2003, December 1). WLAN Testing Reports: “Debunking the Myth of SSID Hiding”

Retrieved April 6, 2008, from

http://www.icsalabs.com/icsa/docs/html/communities/WLAN/wp_ssid_hiding.pdf
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 56
04/18/08

Ou,G. (2008, March 11). PCI security standard endangers wireless LANs. ZDNet

Retrieved April 6, 2008, from

http://blogs.zdnet.com/security/?p=941

Pereira, J. (2007, May 4). BREAKING THE CODE. How Credit-Card Data

Went Out Wireless Door. The Wall Street Journal Online.

Retrieved April 4, 2008, from

http://online.wsj.com/article/SB117824446226991797.html

Phifer, L. (2003, June 30).10 Common questions (and answers) on WLAN security.

SearchSecurity.com

Retrieved April 6, 2008, from

http://searchsecurity.techtarget.com/tip/0,289483,sid14_gci912555,00.html

Phifer, L. (2007a, January 29). Getting Phished: Why SSID Spoofing (Still) Matters

Wi-Fi Planet.

Retrieved: April 4, 2008, from

http://www.wi-fiplanet.com/tutorials/article.php/3656661

Phifer, L. (2007b, March 23). WPA PSK Crackers: Loose Lips Sink Ships

Wi-Fi Planet.

Retrieved: April 4, 2008, from

http://www.wi-fiplanet.com/tutorials/article.php/10724_3667586_2
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 57
04/18/08

Phifer, L. (2007c, September 6). WLAN Security Service Aims to Boost PCI Compliance

eSecurityPlanet.com

Retrieved: April 5, 2008, from

http://www.esecurityplanet.com/best_practices/article.php/3698111

Phifer, L. (2007d, October 5). WLAN Security Blamed for TJX Payment Card Breach

Wi-Fi Planet.

Retrieved: April 5, 2008, from

http://www.wi-fiplanet.com/news/article.php/3703636

Phifer, L. (2007e, December 12). The Caffe Latte Attack: How It Works—and How to Block It.

Wi-Fi Planet.

Retrieved: February 21, 2008, from

http://www.wi-fiplanet.com/tutorials/article.php/3716241

Rubens, P. (2007, April 5). The Critical XP Wi-Fi Patch You Need Today.

Enterprise IT Planet.com

Retrieved April 6, 2008, from

http://www.enterpriseitplanet.com/networking/features/article.php/3669941

Shimonski, R. J. (2003, February 24). Wireless Attacks Primer. WindowSecurity.com

Retrieved April 4, 2008, from

http://www.windowsecurity.com/articles/Wireless_Attacks_Primer.html
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 58
04/18/08

Stubblefield, A., Ioannidis, J., & Rubin., A.(2001, August 21) "Using the Fluhrer, Mantin, and Shamir

Attack to Break WEP", ATT Labs Technical Report, TD4ZCPZZ, Revision 2.

Retrieved April 4, 2008, from

http://citeseer.ist.psu.edu/stubblefield01using.html

Tews, E., Pychkine, A., & Weinmann, R-P. (2007a, April 5). aircrack-ptw

Retrieved April 4, 2008, from

http://www.cdc.informatik.tu-darmstadt.de/aircrack-ptw/

Tews, E., Pyshkin, A., & Weinmann, R-P. (2007b, September 16). Breaking 104 bit WEP in less than

60 seconds. Cryptology ePrint Archive, Report 2007/120.

http://eprint.iacr.org/2007/120

Whitman, M. E. & Mattord, H. J. (2004, November 23) Principles of Information Security. Second

Edition. Thompson Course Technology.

Wi-Fi Alliance. (2003, April 29). Wi-Fi Protected Access: Strong, standards-based, interoperable

security for today’s Wi-Fi networks.

Retrieved April 4, 2008, from

http://www.wi-fi.org/files/wp_8_WPA%20Security_4-29-03.pdf
Aaron Sawyer SECURING WIRELESS NETWORKS: CLOSING THE BACK DOOR 59
04/18/08

Wilson, T. (2007, November 15). Many Retailers Open to Wireless Attacks. Dark Reading.

Retrieved April 4, 2008, from

http://www.darkreading.com/document.asp?doc_id=139291

Wirelessdefence.org. (2006). coWPAtty Main Page.

Retrieved April 4, 2008, from

http://www.wirelessdefence.org/Contents/coWPAttyMain.htm

Wong, Stanley. (2003, May 20). The evolution of wireless security in 802.11 networks: WEP, WPA and

802.11 standards. SANS Institute.

Retrieved April 4, 2008, from

http://cnscenter.future.co.kr/resource/hot-topic/wlan/1109.pdf

Wright, J. (2007, April 23). Using PEAP for wireless authentication. Network World

Retrieved April 5, 2008, from

http://www.networkworld.com/columnists/2007/042307-wireless-security.html