As briefly discussed in our last segment on devices, WLANs vary in form and function according to specific needs. It makes sense, then, that WLANs vary in the way communication is achieved and maintained given the different scenarios. As we learned previously, a wireless network exists when RF is transmitted from the access point (AP), establishing a BSA; this is also known as the basic service set, or BSS. An AP begins by transmitting beacons, which advertise the characteristics of the BSS, such as channel, modulation scheme, and protocols supported.
The BSS is the basic building block of the wireless LAN. The BSS refers to the AP connected to the network infrastructure covering a finite area (the BSA) and handles all traffic to and from clients in that area. It may or may not be an infrastructure device, though most stand-alone APs do function as a portal to the wired network, particularly if there is only one AP. The BSS allows a client to wirelessly communicate within the BSA and with other connected clients and network resources.
Expanding the network is accomplished by increasing power output or by the addition of APs. Raising output power brings with it several issues, while the addition of another AP can also create issues, so the design must be considered carefully. Raising output power is not always an option for APs, particularly those designed for the SOHO market. The concerns with raising power to expand the BSA are: first, client radios may be able to hear the access point, but the AP may not be able to hear the client; second, expanding the BSA will allow more client density in the BSA, creating more traffic and possible delays; third, raising power can cause interference with neighboring WLANs, or, as in the 5 gHz band, interference with radar systems at nearby airports.
When adding APs, the first concern is interference. The new AP must not be on the same channel as the first. It must be assigned to a non-overlapping channel. This "co-channel" interference that could occur if APs are assigned the same channel will be discussed in a later segment. A multichannel architecture (MCA) must be perfected in order to allow proper coverage without excessive errors. The addition of another AP creates what is called an extended service set, or ESS. The ESS effectively expands the primary BSA by conjoining its BSA to the first point. The key to an ESS is proper configuration; aside from proper channel assignment, each of the APs in an ESS must also be configured with the same service set identifier (SSID) and password. As shown in the illustration, there must also be some overlap of the adjacent BSS-typically about 30%. This allows for uninterrupted roaming to occur between each BSS. Theoretically, a WLAN can be expanded indefinitely, observing proper channel assignment. However, a practical limit will be reached if an excessive number of "hops" between wireless APs occurs. It should be understood that roaming on an ESS is done through the wired infrastructure. Each new AP added to the ESS must be wired also. On a mesh type network, the backhaul is usually done wirelessly, but the same limitations apply.
In the above scenario, it is presumed that all, or the majority, of APs are infrastructure devices with a wired connection into the network. But what if you don’t have the ability to wire every AP? What if the proposed coverage area is remote and has only power available? There are two ways to handle this problem: point-to-point (p2p) bridging and mesh topology. In the former, two radios are dedicated to the task of establishing a wireless link over a long distance. A good example is where two buildings on a campus are separated by several hundred feet, without any dedicated wiring or fiber between them. Typically, highly directional antennas are used to establish the link.
In a p2p link, one AP is the root node, while the other, or others, are non-root. This means that one node has control of the radio traffic, which avoids routing and quality of service (QoS) issues. Only one AP in a p2p link can be the root. It is possible to have multiple clients on this type of link: the root AP would broadcast over an omni-directional antenna, while the non-root, client nodes would use directional antennas aimed at the root antenna. Designing and implementing this type of network requires an understanding of local zoning laws governing antennas and towers, so don’t overlook local considerations. Another consideration, as you would expect, is the presence of obstacles or electromagnetic interference (EMI) in the path, or radio line of sight (RLOS). Over long distances, the Fresnel Zone must be calculated and plotted on a good quality topographic map; this will show elevations along the path, along with buildings, towers, and airports. However, the Fresnel Zone must not be obstructed by more than 40%, which means a physical survey of the path must be undertaken. For very long links longer than 7 miles, a phenomenon known as "earth bulge" must be considered. For a directional antenna at ground level, earth bulge becomes a problem at about 7 miles from the antenna. Raising the antenna to the top of a building or tower can mitigate this potential problem.
The other topology, mesh configuration, is more complicated. Very few vendors are capable of operating in a mesh configuration. In a mesh network, each AP cooperatively manages and controls the traffic between nodes. At least one node must be wired into the network infrastructure. This node is called the "mesh portal" and is the access point through which all network traffic accesses wired resources. In a properly designed mesh network, you need at least two portals. A single point of failure is not good practice. The book is still being written on best practice, but it is a good rule of thumb to limit four mesh APs to each portal. This can be further broken down to how many hops data needs to take between mesh APs.
While cooperative mesh points do a great job in routing and controlling data, it comes down the simple fact that the mesh APs operate like repeaters between themselves, and particularly for the mesh points that do not have a network association with the portal AP. This would require a client’s traffic with the farthest AP (without portal association) to be forwarded to the nearest (lowest link cost) neighbor, which would forward or repeat the traffic to another portal until it reaches the portal AP. As we saw in the previous segment, repeating a signal comes with a significant cost in both signal strength and quality. Each hop can introduce 3 dB or more loss into the process; after three hops under ideal conditions, the signal may not be usable. Another good rule of thumb is to limit hops to two APs. This requires careful design that allows multiple paths to the mesh portals.
Mesh networks are essentially self-forming and self-healing, making them very robust. A mesh topology is also distributed in terms of control and data handling, making it an excellent fit for use with distributed control systems. Even if the mesh portals are disabled, the network continues to operate, allowing for the transfer of data to continue between nodes. In a DCS, this loss of communication with the servers is not necessarily a problem in the short term; proper design would eliminate any single point of failure, such as a single mesh portal. Another consideration is the use of a WLAN controller in a controller-based network. A controller can also be a single point of failure in an otherwise robust wireless network. The pros and cons of controllers will be discussed in an upcoming segment.
One last network topology in common use is known as an independent basic service set, or IBSS. This type of network is also called an "ad hoc" network. An IBSS does not require an AP to operate. The network exists between mobile wireless devices and is typically set up "on the fly." Personal hotspots are an example of this type of network. This type of network is useful if there are several colleagues wishing to collaborate on a project without being exposed to a common wireless network in the office or being on an insecure public hotpot. Most operating systems allow for the setup of an ad hoc network using a wizard or some other utility.
In an ad hoc network, the SSID is the computer name or BSSID of the first device to begin sending beacons. Clients wishing to join the network would do so as with a BSS with the originating device controlling traffic and security. Caution must be exercised when using this method. If any of the networked computers has a simultaneous wired connection, this could provide an unauthorized path into the secure network environment. Security for an ad hoc network must be very robust, particularly in public venues. Data obtained by eavesdropping on these sessions can be very sensitive and create havoc for the entities involved in the ad hoc session.
– Daniel E. Capano, owner and president, Diversified Technical Services Inc. of Stamford, Conn., is a certified wireless network administrator (CWNA). Edited by Chris Vavra, production editor, CFE Media, Control Engineering, email@example.com.
www.controleng.com/blogs has other wireless tutorials from Capano on the following topics:
Wi-Fi acronyms, wireless buzzwords, WLAN nomenclature, wireless terms
Wi-Fi and the OSI model
www.controleng.com/webcasts has wireless webcasts, some for PDH credit.
Control Engineering has a wireless page.