Network Topology and Frequency Planning Interference aspects may severely limit the number of links in a network if appropriate

caution is not exercised in the earlier stages of frequency planning (see Figure 5.5). In what follows, some general aspects, based on former sections, are illustrated. 1. Since paths of a chain have very sharp angles, using the same channels by changing polarization (HP/VP) may be a good alternative to using two alternate channels in the chain. 2. Figure 5.5 shows the same channel used alternately with horizontal (HP) and vertical (VP) polarization. Upper (U) and lower (L) duplex halves for the transmitters are illustrated in each site. 3. In the tree configuration, and for sharp angles, polarization discrimination ensures the possibility of using the same channel with different polarization (HP and VP). Both transmitters on the common node have the same duplex half (U).

4. In the ring configuration, the same channel, with the same polarization, is employed in the perpendicular paths but with different polarization in the parallel paths. The transmitters are alternately labeled upper (U) and lower (L) duplex halves. Although the picture does not represent a physical ring configuration, the logical configuration and traffic flow are indeed ring in nature. If the ring consisted of an odd number of sites, there would be a conflict of duplex halves, and changing the frequency band would be a good alternative. 5. In the star configuration, as noted earlier, all transmitters on the common node should have the same duplex half (L). Keep in mind that this configuration displays a difficult frequency planning scenario and is very sensitive to the geometry (mutual angles). If the node is a concentration point for high-capacity links, wide bandwidth is required, thus making the allocation of smaller channels in other portions of the network quite complicated. It is recommended that the link carrying the traffic out of the hub should use a frequency band other than the one employed inside the cluster. 6. Mesh configuration presents a complicated frequency planning scenario as a result of several conflicts of duplex halves. In addition, it normally requires more channels than do other configurations. Low Frequency Bands and the FCC Loading Rules The low frequency bands available for high capacity radios in North America are shown in Table 5.1. High capacity radios are also available in the 15, 18, and 23 GHz bands. These radios normally use a single RF channel. Maximum path lengths are limited to five miles or less in the high frequency bands due to rain outage.

The preferred bands in the U.S. for high-capacity microwave systems are the lower 6 GHz (L6) band (5.9256.425 GHz) and the 11 GHz band (10.711.7 GHz). High-capacity microwave radios typically operate at data rates of 135 to 155 Mbps within a 30 MHz channel bandwidth. FCC Part 101.141(a)(3) defines the minimum loading rules for digital microwave radios. Radios using a 10 MHz channel bandwidth or greater must have 50 percent loading within 2.5 years. This requirement applies to all high capacity 1DS3, 3DS3, and OC3 radios in the lower 6 GHz and 11 GHz bands. A radio is considered 50 percent loaded if at least 50 percent of its DS1 channel capacity is being used. A DS1 channel is considered used if it is connected to a DS1/DS0 multiplexer (e.g., a channel bank). There are no DS0 loading requirements. For non-DS0 services, the next largest DS1 equivalent is considered to determine loading. Systems carrying more than 50 percent digital video are exempt from the loading rules. The lower 6 GHz band is increasingly congested. Furthermore, the L6 GHz band has a large number of licensed satellite Earth stations with each Earth station routinely coordinated for the entire 5,9256,425 MHz band, and for the entire geosynchronous arc, even if the Earth station actually communicates with only one transponder on one satellite. In U.S., the congestion in the L6 GHz band has led a number of applicants to seek licenses to operate in the upper 6 GHz band (6,5256,875 MHz) pursuant to waivers that permit them to operate fixed stations in U6 with bandwidths that are greater than the authorized 10 MHz. These waivers issued by FCC were granted upon showing that there were no channels available in the L6 GHz band, that other higher frequency bands were not suitable for the proposed path, and that there were no other alternatives. TABLE 5.1 High-Capacity Frequency Bands Frequency [GHz] 5.925 6.425 6.525 6.875 6.425 6.930 6.875 7.125 7.125 8.500 7.125 8.275 10.7 11.7 Maximum Bandwidth 30 MHz 10 MHz 30 MHz 25 MHz 30 MHz 40 MHz 40 MHz Frequency Plan FCC Part 101 & Canada FCC Part 101 Industry Canada FCC Part 74 US Federal Government Industry Canada FCC Part 101 & Canada

Microwave Systems Engineering System Documentation In addition to the strictly microwave path engineering part of the project, there is a system engineering portion. System engineers usually provide technical direction and design to guarantee overall system integrity by verifying that all subsystems and contractor-furnished equipment are compatible, and that the desired performance is realized. Systems engineering will also provide transport traffic design and complete system integration as well as the network management system (NMS) and its integration into the other MW and fiberoptic systems. In the next phase, application engineers review and translate the entire system configuration requirements into specific hardware implementations, including standardized interface levels, intra- and inter-rack cabling, and original equipment manufacturer (OEM) integration requirements, and they produce the following system documentation: Criteria, methods, standards, and procedures used for MW path engineering A system block diagram that details the main equipment provided by the contractor; all equipment grouped per site with interconnectivity between sites identified A block and level diagram that shows all the equipment provided by the contractor; may also show connections, interfaces, and signal levels to existing equipment at the same site A T1/E1 plan showing the system traffic routing; shows each T1/E1 connection from the originating site to the destination site A channeling plan A synchronization plan

Path engineering results A power consumption document that provides the value of the total power requirements of all equipment provided by the contractor according to the enclosed equipment spreadsheet, on a persite-basis An equipment list (bill of materials) that includes all the equipment that needs to be provided n Site plan A tower profile showing all the radio equipment and transmission lines installed on the tower (see Figure 5.6) Floor plans and equipment layout A rack profile that shows the equipment mounting position on the rack Wiring diagrams showing equipment and inter-rack wiring, cabling, waveguides runs, and so forth Geographical system layout

Equipment Availability Calculations A microwave link can become unavailable for a number of reasons, but this calculation includes only predictable equipment failures. Therefore, it excludes problems caused by misaligned or failed antenna feeder systems, extended loss of primary power, path propagation outages, human error, and other catastrophic events.

Short-term (<10 sec) propagation outages are applied to the performance (not availability) objective and will not be used here. It is important to define the terms and parameters used in equipment availability calculations as follows: A = 100 (1 U) [%] (5.1) where A = availability (percentage of time, percent) U = unavailability (percentage of time, percent) For n pieces of equipment connected in series (tandem), U = U1 + U2 + + Un (5.2) For two pieces of equipment connected in parallel, U = U1U2 (5.3) Unavailability of the microwave radio terminal can be expressed as:

(5.4)

where MTBF = mean time between failure (MTBF = 109/FITS) FITS = failures in time (in 109 hrs) MTTR = mean time to repair MTTR = RT + TT + (1 P)TR (5.5) where RT = repair time on site TT = travel time to the site P = probability that a spare module is available when needed TR = time to obtain the spare module (assume 24 hrs) FITS (failures in time) is an internationally used unit for measuring or specifying failure rates. Because individual components or subsystems are generally highly reliable in their own right, the convention has arisen of using a period of 109 hrs (114,155 yrs) as a time unit or time scale on which to quantify failure rates (or conversely MTTFs); a failure rate of one failure in 109 hrs equals 1 FIT.

Example: The typical protected MW terminal (1 + 1) has MTBF of 2,200,000 hrs; it takes 0.5 hrs to do the actual repair at the remote site, and the travel time is 3 hrs. With good maintenance practice and spare parts available, we can assume P = 95 % (or 0.95). Let us calculate unavailability of four MW hops connected in tandem (daisy-chain). MTTR = 0.5 + 3 + (1 0.95)24 = 4.5 hr Microwave hop (excluding all other equipment) has two terminals in series, so the unavailability is [ For the four-hop system, total unavailability is UTOT = 4 0.000004091 = 0.000016364 Total availability is ( ) ( ) ]

It is important to notice that this number includes only microwave terminals, and all the other equipment is excluded. Calculations that are more detailed should include channel banks and multiplexers, power supplies, and other items. Total unavailability of the microwave link can then be calculated as a sum of the equipment unavailability and the unavailability due to the propagation issues (rain), i.e., path unavailability Availability of Different Network Topologies The question that very often engineers have to answer is related to the selection of the best and most reliable network topology. Each topology has its advantages and disadvantages, but here we will analyze them from the prospective of the individual microwave path availability/reliability. Unavailability could be caused by hardware failures or propagation problems due to rain, or it could be a combination of both. Multipath, under normal propagation conditions, typically does not cause traffic outage and therefore does not contribute to unavailability. So, let us for a moment assume that all the link unavailability values are the same, i.e., U, and calculate and compare average unavailability per cell site for three different network topologies. Shown here is a simplified method of calculation that assumes uncorrelated failures and availability of the switching mechanism (in the ring topology) of 100 percent. When we say a transport network is 100 percent restorable, we generally mean that it has a sufficient amount and distribution of spare capacity so that any single failure can be withstood without service outage. In principle only dual failures can then cause any service outage. There are increasing numbers of mission critical services calling for as little as 30 seconds (or less) of unavailability per year. This may actually require the ability to withstand certain types of dual failures and motivates analysis of the effects of dual failures on single failure restorable designs. Dual failure analysis is beyond the scope of this book and will not be discussed here. Linear and Star/Hub Topology The following calculation, based on Figure 5.7, shows an average improvement of 40 percent in BTS availability (or reduction in unavailability) for the star/hub network topology over linear (daisy-chaining) topology.

Linear (Daisy-chain or tandem) Topology UTOT = U6 + U7 + U8 + U9 + U10 UTOT = U + 2U + 3U + 4U + 5U = 15U UBTS = 15U/5 = 3U (per BTS) Star/Hub Topology UTOT = U1 + U2 + U3 + U4 + U5

UTOT = U + 2U + 2U + 2U + 2U = 9U UBTS = 9U/5 = 1.8U (per BTS)

5.5.3.2 Ring Topology For a larger transmission network it is recommended to use ring configuration (Figure 5.8) as a high-capacity microwave backbone carrying traffic to the switch location. UTOT = U1 (U2 + U3 + U4 + U5 + U6) + (U1 + U2) (U3 + U4 + U5 + U6) + + (U1 + U2 + U3) (U4 + U5 + U6) + (U1 + U2 + U3 + U4) (U5 + U6) + + (U1 + U2 + U3 + U4 + U5) U6

UTOT = U 5U + 2U 4U + 3U 3U + 4U 2U + 5U U UTOT = 5U2 + 8U2 + 9U2 + 8U2 + 5U2 UTOT = 35U2 UBTS = 35U2/5 = 7U2 (per BTS)

If we assume that the microwave link availability is 99.999 percent (0.001 percent unavailability) we can calculate the average unavailability per BTS for three different topologies. For the linear topology of the five cell-site network, the average unavailability per BTS is 0.003 percent (availability = 99.997 percent); for the star/hub topology, the average unavailability per BTS is 0.0018 percent (availability = 99.9982 percent), and for the ring topology, 0.00000007 percent (availability = 99.99999993 percent). Although this is just an illustration, the advantage of using ring topology from the increased availability prospective is obvious.

Let us now consider a secondary ring connected to a node in the primary ring (see Figure 5.9). The secondary ring has links with lower capacity than the primary ring, so the traffic can only flow from the secondary to the primary ring. The total unavailability of a specific node in the secondary ring is calculated by adding the unavailability of the node in the primary ring, cell-site 3, to the unavailability calculated at that specific node in the secondary ring. The unavailability of a specific node in the secondary ring is obtained similarly as in the primary ring. The unavailability of more secondary rings can be calculated in the same way.

For example, if we calculate the unavailability of the site furthest away from the MW Hub, cell-site 7, we would get the following: UBTS7 = UBTS3 + (U6 + U7)(U8 + U9) = (U1 + U2 + U3)(U4 + U5 + U6) +(U6 + U7)(U8 + U9) Again, for the sake of simplicity, we will assume that all the unavailabilities are the same (this will never be the case in the real networks, but the calculation principles will be the same,) i.e., U:

UBTS3 = 3U 3U = 9U2 UBTS7 = 3U 3U + 2U 2U = 9U2 + 4U2 = 13U2 So, if U = 0.001%, UBTS3 = 0.00000009%, ABTS3 = 99.99999991%, and UBTS7 = 0.00000013%, ABTS7 = 99.99999987%. Important: Note that percentages can be added and/or subtracted directly but cannot (a common mistake) be multiplied or raised to the power without prior conversion into the decimal number. Here is the example of detailed calculation: ( )

Connecting cell-site 6 to cell-site 2 instead of to cell-site 3 is even a better solution, and would create so-called dual-homed secondary ring.