Vendor-Specific Topology Extensions

The vendor-specific topology extensions are an enablement of additional network functionality by way of vendor-defined protocols, devices, and topologies. In this section you will learn how workgroup bridges, wireless repeaters, outdoor wireless bridges, and wire- less mesh networks through the use of wireless controllers can enhance the functionality and capability of your wireless deployment.

Workgroup Bridges
You will most likely have times when you have an isolated network that needs access to the rest of the network for access to the server farm and the Internet. You might not be able to run an Ethernet cable to the isolated network, or you might not own the property so you can’t drill holes in the walls, and so on. In this scenario, you would use a WGB topology such as the one shown in Figure 4-6.

Notice that the WGB is used to bridge a wired network to an AP that connects to a distri- bution system.

Cisco offers two types of workgroup bridges:

• Autonomous Workgroup Bridge (aWGB): The aWGB was originally just called a workgroup bridge, but Cisco later changed the name when it introduced the Univer- sal WGB. The aWGB is supported in IOS AP version 12.4(3G)JA and later. The aWGB connects only to upstream Cisco APs, and the AP sees multiple Ethernet clients.
  

• Universal Workgroup Bridge (uWGB): The uWGB is supported on IOS AP version 12.4(11)XJ and later. It allows bridging to upstream non-Cisco APs and appears as a single client.

Repeaters
Recall that in an Extended Service Set (ESS), multiple APs connect clients. This is all well and good until you have clients roaming about who get into areas where coverage is neces- sary but not possible. The solution of a WGB doesn’t work, because a WGB connects users who are wired. An example is a worker at a warehouse who carries a barcode scan- ner or even a wireless Cisco IP Phone. There are scenarios where you can’t run a cable into a location to install an AP. This is where you want to use a wireless repeater. A wireless repeater is simply an AP that doesn’t connect to a wired network for its connectivity to the distribution network. Instead, it overlaps with an AP that does physically connect to the distribution network. The overlap needs to be 50 percent for optimal performance. Figure 4-7 shows an example. A repeater is allowing a client to connect to the network when in fact the client would normally be out of the service area of the AP.

You can get APs that act as a repeater as well, which is how the Cisco solution works. The catch is that you need a Cisco AP as the upstream “root” device, and only one SSID is supported in repeater mode. Additionally, the overall throughput is cut in half for each re- peater hop.


Outdoor Wireless Bridges

When you have two or more LANs within a few miles of each other and you want to link them, you can use a wireless bridge. Because you are “bridging,” the technology works at Layer 2. This means that the LANs do not route traffic and do not have a routing table.

You can connect one LAN directly to another in a point-to-point configuration, as shown in Figure 4-8, or you can connect many LANs through a central hub, as shown in Figure 4-9.


Each end of a point-to-multipoint topology would have to communicate through the hub if it wanted to communicate with the others. Cisco offers the Cisco Aironet 1300 series wireless bridge and the Cisco Aironet 1400 series wireless bridge. When using a 1400 se- ries, you can bridge only networks, but if you use a 1300 series, you can allow clients to connect as well as bridge networks. The 1300 series operates in the 2.4-GHz range, and the 1400 series operates in the 5-GHz range.


Outdoor Mesh Networks

As you can see, bridges are a good way to connect remote sites. However, suppose that you are operating in a point-to-multipoint topology, and the central site experiences con- gestion. Who suffers? Just the central site? Just the remote site? No; the answer is every- one. When two remote sites communicate through a central site, the central site makes all the difference.

Assume that the central site goes down, as shown in Figure 4-10.


Now the remote sites can’t communicate with each other or the central site. This can be a major issue to contend with. The solution is to deploy a mesh network such as the one illustrated in Figure 4-11.

The mesh solution is appropriate when connectivity is important, because multiple paths can be used. The IEEE is currently working on a mesh standard (802.11s). However, the solution discussed here is a Cisco solution in which a wireless controller, also shown in Figure 4-11, is involved.

When you have a mesh network, some nodes (another term for APs in a mesh network) are connected to a wired network. Some nodes simply act as repeaters. A mesh node re- peats data to nearby nodes. More than one path is available, so a special algorithm is used to determine the best path. The alternative paths can be used when there is congestion or when a wireless mesh node goes down.

Original 802.11 Topologies

Although the previous sections discussed network topologies that you might encounter, it was a very general discussion. You also need to understand the original topologies, defined by the 802.11 committees, including the following:

• Ad hoc mode
• Infrastructure mode

The following sections give more details on these topologies.

Overview of Ad Hoc Networks

When two computers want to communicate directly with one another, they do so in the form of an ad hoc network. Ad hoc networks don’t require a central device to allow them to communicate. Rather, one device sets a group name and radio parameters, and the other uses it to connect. This is called a Basic Service Set (BSS), which defines the area in which a device is reachable. Because the two machines don’t need a central device to speak to each other, it is called an Independent Basic Service Set (IBSS). This type of ad hoc network exists as soonas two devices see each other. Figure4-2 shows an ad hoc network.

Each computer has only one radio. Because there is only one radio, the throughput is lower and acts as a half-duplex device, because you can’t send and receive at the same time.

You don’t have much control in these networks, so you’re stuck when it comes to methods such as authentication. In addition, you need to address who starts the conversation and who decides on the order of communication, to name just a couple issues.

Network Infrastructure Mode

In wireless networks, an access point acts as a connection point for clients. An AP is actually a cross between a hub and a bridge. Here’s why:

• There is one radio, which cannot send and receive at the same time. This is where the AP is likened to a hub. It’s a half-duplex operation.
• APs have some intelligence that is similar to that of a bridge. That is how an AP can see a frame and decide to forward it based on MAC addresses.

What is different on an AP versus a bridge is that wireless frames are more complex. Standard Ethernet frames have a source MAC address and a destination MAC address. Wireless frames can have three or four MAC addresses. Two of them are the source and destination MAC addresses, and one is the AP’s MAC address that is tied to a workgroup.


The fourth that could be present is a NEXT_HOP address in the event that you are using a workgroup bridge (WGB).

An AP is actually just one type of wireless station. This terminology could cause some confusion between an AP and a client on a network, so to differentiate between them, a client is called a station (STA), and an AP is called an infrastructure device.

So what does a typical wireless topology look like? Of course, wireless clients are associated with an AP. In the wireless space, the coverage area of the AP is called a Basic Service Area (BSA), which is also sometimes known as a wireless cell. They mean the same thing. When only one AP exists, this coverage area is called a BSA, as shown in Figure 4-3. That AP then usually has an Ethernet connection to an 802.3 LAN, depending on the function of the AP.

Assuming that the AP has an Ethernet connection, it bridges the 802.11 wireless traffic from the wireless clients to the 802.3 wired network on the Ethernet side.

The wired network attached to the AP’s Ethernet port is a path to a wireless LAN controller (or controller for short). The client traffic is passed through the controller and then is forwarded to the wired network, called the distribution system. The distribution system is how a client accesses the Internet, file servers, printers, and anything else available on the wired network.
When more than one AP is connected to a common distribution system, as shown in Figure 4-4, the coverage area is called an Extended Service Area (ESA).


Why would you want more than one AP connected to the same LAN? There are a few reasons:

• To provide adequate coverage in a larger area.
• To allow clients to move from one AP to the other and still be on the same LAN.
• To provide more saturation of APs, resulting in more bandwidth per user.

This process of a client moving from one AP to another is called roaming. For roaming to work, the APs must overlap. You might wonder why they need to overlap, because interference in a wireless network is a common issue. The reason for the overlap is so that a client can see both APs and associate to the one with the stronger signal. As soon as the signal from the associated AP hits the threshold built into the client, the client looks for another AP with a better signal.


Service Set Identifiers

Think about how you connect to a wireless network. On your laptop, you might see a popup that says “Wireless networks are available” or something to that effect. When you look at the available networks, you see names. On older Cisco autonomous APs, the network was called “Tsunami.” On a store-bought Linksys, the network is actually called “linksys.” So the client sees a name that represents a network.

On the AP, the network is associated with a MAC address. This network or workgroup that your clients connect to is called a Service Set Identifier (SSID). So on an AP, the SSID is a combination of MAC address and network name. This MAC address can be that of the wireless radio or another MAC address generated on the AP. When an AP offers service for only one network, it is called a Basic Service Set Identifier (BSSID). APs offer the ability to use more than one SSID. This would let you offer a Guest Network and a Corporate Network and still use the same AP. When the AP has more than one network, it is called a Multiple Basic Service Set Identifier (MBSSID). You can think of it as a virtual AP. It offers service for multiple networks, but it’s the same hardware. Because it’s the same hardware and the same frequency range, users on one network share with users on another and can collide if they send at the same time.

Now let’s return to the roaming discussion. To get roaming to work, the BSA of each AP must overlap. The APs also need to be configured for the same SSID. This enables the client to see that the same network is offered by different MAC addresses, as illustrated in Figure 4-5.


When a client roams and moves from one AP to the other, the SSID remains the same, but the MAC address changes to the new AP with a better signal.

Another issue to consider when roaming is the possibility of interference between the two overlapping APs. Even though they offer the same SSID, they need to be on different channels, or frequency ranges, that do not overlap. This prevents co-channel interference, which should be avoided. The 2.4 spectrum allows only three nonoverlapping channels. You must consider this fact when placing APs.

General Wireless Topologies

When you’re talking about wireless topologies, there are a number of ways it could go. If you are talking about how your wireless network looks next to your wired network, you are most likely talking about a wireless local-area network (WLAN). The goal of a WLAN versus a wireless personal-area network (WPAN) is quite different. The following sections discuss the purpose of each network type, what they try to accomplish, and what types of wireless technologies you might encounter there. Figure 4-1 shows the various wireless topologies.

WPAN

If you were to consider all the options, a WPAN would be the solution to choose if you wanted to wirelessly connect to something that is very close to you. It seems funny to put it that way, because if something close to you needs to be networked, you might as well just walk over and grab it, right? Wrong. Even though this is called a network, its form can mislead you into thinking that it’s not a networking technology. What forms are we talking about? Headsets, headphones—even a mouse.
A WPAN has the following characteristics:

• The range is short—about 20 feet.
• Eight active devices
• Unlicensed 2.4-GHz spectrum
• Called a piconet

A WPAN is a network that is designed to operate within a 20-foot range. The most common WPAN is Bluetooth. In a Bluetooth network, you communicate on the 2.4-GHz spec- trum. Thinking about how many people have Bluetooth headsets and mice and such, you would expect a lot of interference, but that’s not the case. Bluetooth uses Frequency Hop- ping Spread Spectrum (FHSS). Although this book doesn’t discuss FHSS, it’s good to un- derstand that even though Bluetooth operates on the same frequency as 802.11b and 802.11g, they don’t interfere as much as another AP in the same frequency spectrum would, but they do interfere. The fact that Bluetooth communicates with a shared hopping sequence in a local area is what makes it a piconet.

Bluetooth piconets consist of up to eight active devices but can have many inactive de- vices. WPANs usually fall into the unlicensed 2.4-GHz spectrum and are standardized by the 802.15 IEEE workgroup. A WPAN study group was formed in 1998, and two months later a Bluetooth Special Interest Group (SIG) was formed. Shortly thereafter the study group became the IEEE 802.15 group. The Bluetooth SIG has more than 9000 members and continues to further the technology.


WLAN

WLANs are designed for a larger area than that of a WPAN. These can scale from very small home offices to large enterprise networks. The fact that they are local-area means that the organization where the WLAN exists also manages and probably owns the equip- ment. WLANs have the following characteristics:

• 2.4-GHz or 5-GHz spectrum.
• A larger range than a WPAN—close to 100 meters from AP to client.
• To achieve further distance, more power output is required.
• It’s not personal; rather, more clients are expected.
• WLANs are very flexible, so more than eight active devices/clients are expected, un- like a WPAN.

Normally you find a mix of dual-band wireless access points, laptops, and desktops in a WLAN. A WLAN operates in either the 2.4-GHz spectrum for 802.11b/g or the 5-GHz spectrum for 802.11a. Of the protocols seen in WLANs, 802.11b was the first to really get market penetration. Others, such as the 802.11a, have followed. Now the 802.11a, b, g, and n WLAN standards are commonly found in networks around the world. The frequency spectrums used by 802.11a/b, g, and n are all unlicensed.

Because WLANs cover larger areas, they require more power output than a WPAN. The issue to watch in WLANs is that you don’t exceed the power rules that the government sets forth. For example, in the U.S., the Federal Communications Commission (FCC) man- dates radiated power levels.

WLANs are designed to give mobile clients access to network resources. For this reason, a WLAN expects to see multiple users. In addition to wireless users, there are wireless print servers, presentation servers, and storage devices. You end up with many devices connecting to each other or sharing information with each other, usually over a common distribution system such as the local-area network. This makes WLANs much more com- plex than WPANs.

What makes WLANs flexible is the fact that the APs and clients are dual-band. This makes it easy to deploy different transmission methods in different areas, and most clients can still operate.


WMAN

A wireless metropolitan-area network (WMAN) covers a large geographic area and has the following characteristics:

• Speeds decrease as the distance increases.
• Close to broadband speeds versus Ethernet speeds.
• Used as a backbone, point-to-point, or point-to-multipoint.
• Most well-known is WiMax.

WMANs are used as backbone services, point-to-point, or even point-to-multipoint links that can be a replacement for technologies such as T1 and T3. Sometimes, a WMAN can use unlicensed frequencies. However, this isn’t always a preferred solution, because others could use the same frequency, thus causing interference. Instead, many prefer to use a li- censed frequency range; however, this requires payment for exclusive rights.

It’s normal for the speeds in a WMAN to decrease with distance. This places them in a closer category to broadband than to Ethernet. The most widely known WMAN is WiMax (802.16b). WiMax can be used to offer last-mile access as an alternative to broad- band services such as DSL or cable connections. WiMax is an excellent solution where fa- cilities or distance are a limitation. With WiMax, you pay a service provider for access, because the cost of deployment is normally very high.


WWAN

A wireless wide-area network (WWAN) covers a large geographic area. WWANs have the following characteristics:

• Low data rates
• Pay-for-use
• High cost of deployment

Because they cover a large geographic area, WWANs usually are very expensive to deploy. To better understand what a WWAN is, consider your cellular service. Your cell serv- ice is a WWAN and probably offers data access as well as voice access. The data rates are probably around 115 kbps, although some providers offer higher data rates. The most widely deployed WWAN technologies are Global System for Mobile Communication (GSM) and Code Division Multiple Access (CDMA). Payment for data access or even voice access is typically based on usage.