WLAN requirements generally call the ability of users to roam with wireless client devices throughout the coverage area. Each client radio will roam differently, depending on proprietary protocols that the vendor has incorporated into the radio. The type of wireless application that the client device is using also impacts the ability to roam. As a result, it is crucial that you test the roaming capability of all client device before finalizing the design of the WLAN.
When designing a WLAN, consider the following elements that impact roaming:
- Roaming levels
- Wireless IP phone roaming
- Mobility settings
As a basis for designing a WLAN that provides effective roaming, it is important to first understand basic roaming concepts. For WLANs, roaming takes place at several levels. As shown in Figure 11-14, roaming can take place at Layer 2 or Layer 3. As a user moves about the coverage area, the client radio automatically hands off from one access point to another as needed to support communications. This is Layer 2 roaming, sometimes referred to as access point roaming. In addition, the client device might need to roam from one subnet to another, which is Layer 3 roaming.
Figure 11-14 Layer 2 Versus Layer 3 Roaming
Access Point Roaming
Through the collaboration of WLAN vendors, the Inter–Access Point Protocol (IAPP) specification provides a common roaming protocol enabling users (client radios) to move throughout a facility while maintaining a connection to the network via multivendor access points. Interoperability tests and demonstrations show that IAPP works with a variety of access points. The Wi-Fi Alliance includes interoperable roaming as a requirement to receiving Wi-Fi certification.
The IAPP specification builds upon the capabilities of the IEEE 802.11 standard, using the distribution system interfaces of the access point that 802.11 provides. IAPP operates between access points, using the User Datagram Protocol (UDP) and the Internet Protocol (IP) as a basis for communications. IAPP defines two basic protocols: the Announce Protocol and the Handover Protocol. The Announce Protocol provides coordination between access points by performing the following functions:
- Informs other access points about a new active access point
- Informs the access points of networkwide configuration information
The Handover Protocol informs an access point that one of its stations has reassociated with a different access point. The old access point forwards buffered frames for the station to the new access point. The new access point then updates filter tables to ensure that MAC-level filtering (bridging) will forward frames appropriately.
The finalization and proliferation of IEEE 802.11r and 802.11k amendments to the 802.11 standard have positive impacts on access point roaming. 802.11r provides seamless roaming between access points. The main application for 802.11r is for providing effective roaming for VoIP and security mechanisms. 802.11r provides functions for determining QoS and performing security protocol handshakes before handoffs occur to avoid delays after handoff. 802.11k works in conjunction with 802.11r by providing information to discover the best available access point for handoff purposes. Consider incorporating 802.11r and 802.11k into VoIP and security applications.
As a wireless client device roams from one IP subnet to another, the client device might need to obtain a valid IP address for the new subnet. The client device can make use of DHCP to obtain the IP address, but this is not always effective when supporting mobility. DHCP is not designed to renew addresses when crossing subnet boundaries. As a result, it might be necessary to configure a WLAN to operate on a single subnet.
This might work in a private network, but the subnet roaming issue resurfaces when the client device needs to roam to another network. Consider the use of wireless middleware for applications that are affected by subnet roaming issues.
Some companies, however, might want to implement multiple IP subnets across a common WLAN for various reasons, such as to make network management easier, facilitate location-based services, and decrease the spread of broadcast packets throughout the network. For instance, a company might want to deliver specific information to users based on their location in a specific building. By designating different subnets throughout the WLAN, the location of the user can be found and content delivered to the user based on their location. The location of the user in an airport, for example, could deliver a map of the applicable concourse or terminal, indicating flight information and the location of coffee shops and ticket counters. The system could also deliver advertisements from concessions located within the general area.
With multiple subnets mobile users must be able to seamlessly roam from one subnet to another while traversing a facility. As users roam across subnets, though, there must be a mechanism at Layer 3 to ensure that the user device configured with a specific IP address can continue communications with applications. Some controller-based WLAN implementations, such as Cisco using Layer 3 services, however, can solve this via a feature called Proxy ARP, where a wireless client device can roam from an initial anchor controller to another foreign controller while maintaining the IP address originally assigned via the anchor controller. This is possible even though the foreign controller is operating on a different subnet than the anchor controller.
If the wireless solution you choose does not implement Proxy ARP, consider Mobile IP, which solves Layer 3 roaming by allowing the mobile user to use two IP addresses. One address, the home address, is static. The second address, the "care-of" address, changes at each new point of network attachment and can be thought of as the user's position-specific address.
The home address enables the mobile node to continually receive data relative to its home network, through the use of a network node called the home agent. Whenever the user is not attached to the home network, the home agent receives all the packets sent to the mobile user and arranges to deliver them to the mobile user's current point of attachment, which is its care-of address. Whenever a Mobile IP user moves, it registers its new care-of address with its home agent. This makes it possible for the home agent to keep up with the whereabouts of the mobile user. The home agent then sends any packets it receives for that user to the applicable care-of address.
To implement Mobile IP, you need two major components: a Mobile IP server and Mobile IP client software. The Mobile IP server will fully implement the Mobile IP home agent functionality, providing mobility management for the mobile users. The Mobile IP server can generally also keep track of where, when, and how long users use the roaming service. That data can then provide the basis for accounting and billing purposes.
The requirement for client-side software makes Mobile IP impractical for some applications. For example, public networks demand open connectivity for users, which makes it difficult to deploy solutions that require client software. Of course, the task of installing software on user devices before enabling roaming is too cumbersome. Another problem with Mobile IP is that it is somewhat vendor specific. To ensure interoperability among multivendor Mobile IP clients and servers, definitely do some upfront testing.
Wireless ISP Roaming
With Wi-Fi hotspots, there is very limited roaming among wireless ISPs. The Wi-Fi Alliance had tried developing standards several years ago to make wireless ISP roaming seamless, but the group later disbanded due to significant incompatibility among differing access controllers. In general, standards for roaming from one Wi-Fi wireless ISP to another are nonexistent. As a result, you must negotiate teaming agreements with other ISPs when deploying widespread Wi-Fi hotspots and create a custom access control system that is common among all applicable ISPs.
Wireless IP Phone Roaming
One the applications that roaming impacts a great deal are wireless telephony. If you plan to have wireless IP phones operating on the WLAN, ensure that the WLAN you design supports effective roaming. The voice user must be able to move about the facility, and the system will need to allow roaming at both Layer 2 and Layer 3. Without smooth roaming, users will experience dropped calls.
Layer 2 roaming takes place when a wireless IP phone moves out of range of an access point and reassociates with a different access point. Because this type of roaming will occur frequently as users move about a facility, such as a warehouse or hospital, be certain that Layer 2 roaming is fast (ideally less than 100 milliseconds). This mid-call roam time is the amount of time that elapses after the last RTP packet is seen from the current access point and the first RTP packet seen from the access point that the wireless IP phone associates with.
When selecting wireless IP phones, be certain that they support fast roaming. The Cisco 7920 wireless IP phones, for instance, are specially designed to make Layer 2 roaming fast enough to avoid dropped calls. For example, a Cisco 7925 phone will initiate a reassociation process with a different access point if the phone does not receive three consecutive beacons from the existing access point and its unicast frame to the access point is not acknowledged. The 7920 also periodically scans for better access points and maintains a list of potential access points. A decision to roam is made on signal strength and signal quality metrics. The quality metric makes use of information provided by each access point in its beacon regarding the utilization of the access point. As a result, the phone can avoid attempting to reassociate with access points that have high utilization and may not be able to effectively support voice traffic. With these mechanisms, the 7925 is generally able to complete Layer 2 roaming within 100 milliseconds.
The addition of Layer 3 roaming might cause substantial delays that may drop voice calls. This will depend on the wireless solution that you implement, especially those that do not implement enhanced features for voice roaming. As a result, you might need to consider avoiding having wireless IP phones perform mid-call Layer 3 roaming. In this case, if possible, define a single subnet for the entire WLAN. This will completely avoid Layer 3 roaming. If this is not possible or feasible, at least minimize the possibility of Layer 3 mid-call roaming. For example, you may have a different subnet on each floor of a hospital. When performing the wireless site survey, ensure that signals from one floor do not overlap at such an extent that the wireless IP phones will roam to the floor above (or below) while the user is walking through a particular area.
Many WLAN product vendors enable you to indicate the degree of mobility of each station so that the access point can optimize roaming algorithms. If you set up the station as being "mobile," the roaming protocols will enable the station to reassociate as it moves from one access point to another. If set to "mobile" mode, stationary devices, such as wireless desktops, might experience a short episode of radio signal interference and falsely reassociate with a different access point. As a result, stationary roaming modes usually take this into consideration.