Point-to-Point Microwave

Point-to-point microwave links are rare in the local loop because operators would require a separate base station unit for each customer. They provide higher capacities than multipoint systems, so are occasionally used to connect customers who would otherwise need fiber, but their main application is for private networks.

The classic example of a private point-to-point network is between skyscrapers in a city. If a company has two offices with a clear line of sight between them, it can save on call costs and leased line charges by installing microwave transceivers on each. This type of setup, illustrated in Figure 11.5, is called a wireless bridge. Typical systems have a range of up to about 20 km, so the offices could be in different states or even different countries. For example, some U.S. companies avoid expensive calls to and from Mexico by setting up a wireless bridge with one node in San Diego and the other in Tijuana.

Figure 11.5. Private point-to-point network (wireless bridge)


These systems can operate in both licensed and unlicensed bands. The most important are listed in Table 11.3

Licensed

Most microwave radio systems operate at frequencies between 18 and 38 GHz, where relatively large slices of bandwidth are available. They generally offer throughput of up to 155 Mbps, the same as LMDS. The advantages are that this capacity is two-way, it is dedicated to a single user rather than shared by an entire cell, and the range is increased by focusing the beam on a specific recipient.

Spectrum between 18 and 38 GHz always requires a license, though some vendors and user groups are pushing for this requirement to be removed; because point-to-point links confine their beams to very narrow area, there is little risk of interference. Fortunately, many governments have adopted rather liberal licensing policies, so in some countries it is little more than a formality.

Different parts of the spectrum are available for point-to-point wireless in different parts of the world. North America and Europe have agreed on four bands: 18, 23, 28, and 38 MHz. Europe has an additional band at 26 MHz, which cannot be used in the U.S. Their characteristics are very similar, and most equipment vendors sell transceivers for all of them; some can even be adapted, depending on which part of the world they are in.

The 28 GHz band is the same as that used for LMDS, meaning that licenses to use it have already been allocated. This may make it seem rather pointless, but some LMDS operators do find it useful. They can use point-to-point links for customers who require higher capacity than LMDs can offer, and also to link together their base stations.

Unlicensed

Point-to-point links can also be set up in unlicensed parts of the spectrum, although their performance will not be as good as their licensed counterparts. One option is the familiar ISM band, using IEEE 802.11 or similar equipment. It offers the same 11 Mbps capacity as in multipoint links, but this is dedicated to one user instead of shared over a cell.

For higher capacity links, the FCC actually has a special band dedicated to unlicensed point-to-point microwave. It occupies spectrum in the 5 GHz region and is known as U-NII (Unlicensed National Information Infrastructure). Using proprietary systems, vendors such as Proxim and Western Multiplex sell equipment that can send and receive at up to 100 Mbps in this band.

The disadvantage of the U-NII band is that it has only been granted by the U.S., not internationally. Some other countries may decide to adopt it, but Europe will not—it has already granted the same spectrum to the HiperLan wireless LAN standard. ETSI is trying to develop a version of HiperLan that can be used for local loop and building-to-building applications, but this isn't ready yet. Known as HiperAccess, it will offer two-way links at 20 Mbps.

Table 11.3. Point-to-Point Microwave Bands in Europe and North America
Band NameActual FrequenciesUsed inMax. CapacityLicensed?
ISM2.4–2.483 GHzBoth11 MbpsNo
U-NII5.15–5.35 GHz 5.725–5.825 GHzU.S.100 MbpsNo
18 GHz17.7–19.7 GHzBoth155 MbpsYes
23 GHz21.2–23.6 GHzBoth155 MbpsYes
26 GHz24.5–26.5 GHzEurope155 MbpsYes
28 GHzSee Figure 11.4 Both155 MbpsYes
38 GHz37–39.5 GHzEurope155 MbpsYes
38 GHz37–38.4 GHz 38.6–40 GHzU.S.155 MbpsYes

Network Architecture

Although point-to-point radio is not usually thought of as ideal for WLL, some operators are using it. Instead of trying to connect every customer to their own central site, these companies connect each customer to another, a process known as daisychaining

The customers can't actually access each other's data or communicate directly (unless they want to); the intention is simply to extend the range of the system. Despite the limit of about 20 km, a customer could be much further away than this, using multiple hops to reach the operator. These extra hops make the system less reliable, as they introduce extra points of failure, so extra redundant links are sometimes added. For example, the operator Advanced Radio Telecom is building several 38 GHz networks around the U.S., with customers connected in a series of interlocking rings.

normal: Rings or Stars?

In general, networks of all types can be arranged in three different ways, known as topologies. Real networks are usually a mixture, but the "ideal" cases serve to illustrate the point. The three topologies are shown in Figure 11.6.

  • Star networks connect every user to a central site, such as a mobile base station. This is the typical setup for most wireless networks, including cellular systems and multipoint fixed wireless. It's also the most common type of LAN, with each computer connected to a hub. Star networks are often hierarchical. In a cellular network, many base stations are connected to an MSC; in a LAN, many hubs may be connected to a switch.

  • Ring networks have no central site, instead connecting each user to another. Rings have the advantages that a network can be built up from many short links and that they are resistant to a single failure—though not to more than one. They are popular among long-distance fiber networks, which may cross an ocean in two separate places or even encircle the entire world.

  • Mesh networks connect every node to many others. A fully-meshed network links each node to every other, but these are rare. The only real fully-meshed networks are Bluetooth PANs, some wireless LANs, and occasionally groups of PMR users. More usual is for extra links to be placed in ring networks, both for redundancy and so that the ring doesn't have to be reordered when new nodes are added. The Internet backbone is the best example of a partially meshed network.


Figure 11.6. Network topologies


..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset
52.14.196.121