Return on Investment (roi) for Multi-Technology son juan Ramiro, Mark Austin and Khalid Hamied

7.3.5. Modeling the Need for Capacity Expansions

Assuming an ideal (i.e. error-free) capacity planning process, the first step is to compute the amount of capacity expansions that need to be carried out in order to serve the forecasted traffic. To make this process simpler, it is assumed that the traffic volume is always growing or at some point stable, but never decreasing. Furthermore, the initial calculations are conducted assuming a perfect capacity planning process, and the impact of manual errors will be introduced at a later stage.
Let us assume that all sectors are equipped with infrastructure to manage one UMTS carrier (in this context, a carrier is a transceiver that allows one sector to manage an independent UMTS carrier). At a certain year i, the amount of additional carriers C(i) as compared with the initial configuration of one carrier per sector can be computed as follows:
(7.7)
n=1
where N is the number of sectors in the network, and P_n(i) is the rate of sectors that need at least the activation of the n^th carrier in order to (partially or fully) cope with the amount of traffic that they have during year i. The calculation of P_n(i) is based on the traffic distributions presented in Section 7.3.4.1 (see Figure 7.9 and Figure 7.10). Note that the reason why the

2 × (Carrier capacity)
Figure 7.9 Computing the amount of extra carriers required for a certain traffic distribution (as compared with the original configuration).

2 × (E1 capacity)
Figure 7.10 Computing the amount of extra E1 links required for a certain traffic distribution (as compared with the original configuration).
aforementioned percentage differs from year to year is that the traffic distribution changes as time goes by. In addition, note that for this simple calculation, no spectrum limitation is assumed (this will be addressed in the Self-Optimization example).
Similarly, if the starting point is that all sectors are equipped with HW to manage M carriers, the structure of the process would be similar, and the resulting equation (defining the number of extra carriers as compared with the starting point) would be the following:
∞
C i( ) = N ∗ ∑ (n − M)⋅P i_n ( ) (7.8)
n M⁼
At this point, it is important to highlight that the calculated number of additional carriers, C(i), that are required for year i represents the extra HW that is required for that year as compared with the initial configuration (i.e. with the equipment that is installed at the beginning of the time period for the cash flow discount calculations, also denoted by C(0)). Therefore, for a given year i, the actual number of new additional carriers A(i) is given by:
A(i) = C(i) − C(i − 1) (7.9)
When evaluating capacity expansions in the Iub interfaces, the mechanism is similar, although some differences need to be considered:

• 
  Since Iub interfaces are physically related to NodeBs, not to sectors, it is assumed that their resources are pooled among sectors in the same NodeB.
  
• 
  Therefore, a certain Iub needs a capacity expansion when the aggregated traffic of its c orresponding NodeB exceeds the currently installed capacity, although some safety margins to make calculations more conservative can be introduced.
  
• 
  When characterizing the traffic per site, there are several options, depending on the local homogeneity of the network. In the typical case of three sectors per NodeB, it can be assumed that the average traffic per NodeB in the busy hour equals three times the average traffic per sector, which implicitly endorses the assumption that loaded sectors tend to be relatively grouped together, whereas less loaded sectors also tend to be locally grouped within areas of low loaded NodeBs. This is the assumption that has been considered in this example, although other distributions can be used if more detailed knowledge is available about the traffic distribution across the network under consideration.
  
• 
  When analyzing the investment in new carriers, only extra, additional carriers were considered every year since, in this context, each investment in extra carrier equipment is a one-off payment. However, in this example, let us assume that the E1 links that provide capacity to Iub interfaces are leased lines. In this case, an annual fee needs to be paid for each and every leased line, no matter whether that line was already active during the previous years or not.