hash stringlengths 32 32 | doc_id stringlengths 7 13 | section stringlengths 3 121 | content stringlengths 0 2.2M |
|---|---|---|---|
314690db30e0da820145d5cd1eef3bc5 | 104 141 | 6.3 Scenario description | Clause 6 pursues the task of planning a bidirectional backhaul connection between two radio sites at a distance of 5,3 km and deployed in a geographical region where rainfall intensities exceed 32 mm/h for 0,01 % of the time in a year (see figure 17 for reference). The expected traffic volume to be transported across the two communication paths is unbalanced, and it is characterized by a peak load = 3,5 Gbit/s and an average value = 1,36 Gbit/s in the direction with the highest demand - both of which are known parameters. Figure 17: Overview of the illustrative backhaul scenario considered in clause 6, with information on the link distance, the rain zone and the main traffic statistical parameters Based on this objective, the selected radio equipment consists of a single-polarization transceiver unit operating in the E-band, featuring a bandwidth of 500 MHz and a maximum output power of 26 dBm, with a maximum gross information rate equal to 4,5 Gbit/s yielding a net throughput greater than the expected peak traffic load . The unit supports an Adaptive Coding and Modulation (ACM) policy, and it is therefore capable of dynamically adjusting the delivered capacities according to the fading conditions along the propagation path. In this scenario, two types of parabolic antennas are assumed to be available: one with a diameter of 30 cm (maximum gain of 45,4 dBi), and another with a diameter of 60 cm (maximum gain of 51,4 dBi). These give rise to three possible radio link configurations: β’ Configuration 1: one E-band unit equipped with 30 cm antenna at both the radio sites; β’ Configuration 2: one E-band unit with 30 cm antenna at the first site and one E-band unit with 60 cm antenna at the second site; β’ Configuration 3: one E-band unit with 60 cm antenna at both the radio sites. The purpose of clause 6 is to define the optimal radio link configuration to ensure compliance with the following target conditions, derived from a New-KPIs-oriented link planning methodology (see Introduction for further details): β’ CIR = 25 Mbit/s available for at least 99,995 % of the time ETSI ETSI TR 104 141 V1.1.1 (2026-03) 32 β’ BTA higher than 99,97 % β’ Gross PIR = 4,5 Gbit/s with at least 5 dB of fade margin It is observed that the BTA constraint expressed above can be interpreted, for instance, as stemming from the application of the apportionment rule outlined in annex A of the ETSI GR mWT 028 [i.1] to the network topology depicted in figure 18, where three radio sites - , and - are connected to a remote aggregation node in a daisy-chain configuration (see annex C for further details on the recommended guidelines for BTA apportionment in generalized network scenarios). Specifically, the selection of the target BTA values for each link illustrated in the diagram ensures that the traffic generated by each of the three radio sites achieves an overall end-to-end BTA (i.e. up to node ) that consistently meets or exceeds the threshold of 99,9 % recommended in [i.1]. Figure 18: Illustrative BTA apportionment for a network topology with three radio sites connected to a remote aggregation point in a daisy-chain configuration Two distinct link planning procedures are outlined in the following, depending on whether an estimate of the complete statistical distribution of the expected traffic demand, in addition to the provided peak and average values ( and , respectively), is available or not. |
314690db30e0da820145d5cd1eef3bc5 | 104 141 | 6.4 Link planning with known traffic distribution | When an estimate of the cumulative distribution function of the link's target traffic demand is available - such as from the employment of the measurement-based methodology described in clause 5 - the planning procedure should follow the general steps specified in table 6. Table 6: Link planning procedure according to the New KPIs methodology with known cumulative distribution function of the link's target traffic demand Initialization 1) For each th radio link configuration, compute the following metric: (19) as the sum of the maximum transmit power (in dBm), the transmit antenna gain (in dB) and the receive antenna gain (in dB) of the available radio equipment; 2) sort the available radio link configurations in ascending order based on the metric , to obtain: 1 2 β― , (20) being the total number of radio link configurations; 3) initialize 1; Link Planning 4) compute the metrics prescribed by the New KPIs methodology considering the th radio link configuration: 4a) derive the PIR fade margin as: , (21) ETSI ETSI TR 104 141 V1.1.1 (2026-03) 33 where is the transmit power of the radio equipment relative to the PIR (in dBm), an are defined as in step 1, is the free-space path loss (in dB), is the attenuation due to atmospheric gases (in dB), and () is the receiver sensitivity threshold relative to the PIR (in dBm); NOTE: Term can be computed according to Recommendation ITU-R P.676-13 [i.5]. 4b) derive the availability () of each th backhaul capacity (with () < () < . . . < ()) according to the well-established methodologies based on the fading prediction models described in Recommendation ITU-R P.530-19 [i.4]; 4c) for all indexes = 1,2, β¦ , , compute the probability < β€ that the link throughput demand lies in the range between backhaul capacities and on the basis of the target traffic cumulative distribution function (assumed known in the present clause) as: < β€ = β( ), (22) being = 0 bit/s the link failure state; 4d) derive the BTA through the formula presented in equation (1), which is reported here for ease of reference and improved readability: = ( < β€ ) Γ () , (23) where terms () and ( () < β€ ()) (for = 1,2, β¦ ) have been obtained in previous steps 4b and 4c, respectively; 5) if the PIR fade margin (), the BTA and the CIR availability derived in step 4 meet the planning targets, then end the procedure and plan the link with the th radio configuration, otherwise increment the counter by 1; 6) if β€ repeat the process starting from step 4, otherwise end the procedure and conclude that no radio link configuration among the available ones can meet the planning targets. NOTE: The initialization phase presented in table 6 (and in following table 8) serves solely as an illustrative example. Its purpose is to organize the candidate radio link configurations in a logical sequence, in order to begin with the simplest (i.e. least costly) option and to progressively test more complex (or expensive) solutions in the subsequent Link Planning steps 4, 5 and 6, until all New KPIs target conditions are satisfied. Depending on the specific context, alternative and more sophisticated sorting criteria can be necessary to accommodate varying strategic needs. ETSI ETSI TR 104 141 V1.1.1 (2026-03) 34 Figure 19: Cumulative distribution function of the link's target traffic demand used to derive the BTA values in table 7 through equations (22) and (23) Table 7 shows the values of the PIR fade margin , the BTA ! and the CIR availability that can be obtained by applying steps 4a through 4d of the methodology in table 6 to each th radio link configuration defined in clause 6.3 (here 1,2,3), assuming that the cumulative distribution function of the link's target traffic demand is the one depicted in figure 19. Based on the reported outcomes, it can be easily deduced that the planning procedure of table 6 would result in the selection of the second radio link configuration ( 2), that delivers the minimum system gain needed to satisfy all the New KPIs target conditions specified in clause 6.3. To provide a broader perspective, the results presented in table 7 are complemented by the plot in figure 20, that illustrates the variation of: i) the PIR fade margin (right y-axis); ii) the BTA calculated according to equation (23) on the basis of the link's target traffic demand distribution shown in figure 19 (left y-axis); iii) the lower bound of the BTA derived through the analytical procedure described in clause 4.3 (which forms the core of the approach discussed in the following clause 6.5) (left y-axis); and iv) the CIR availability (left y-axis) as a function of the combined transmission and reception antenna gains. As a reference, the gain levels ensured by radio link configurations 1, 2 and 3, corresponding to 90,8 dBi, 96,8 dBi and 102,8 dBi, respectively, are also highlighted on the x-axis. Table 7: Key results obtained by applying steps 4a through 4d of the planning procedure illustrated in table 6 to the radio link configurations described in clause 6.3 Radio link configuration PIR fade margin BTA CIR availability 1 9,26 dB 99,964 % 99,995 % 2 15,28 dB 99,978 % 99,997 % 3 21,30 dB 99,986 % 99,998 % ETSI ETSI TR 104 141 V1.1.1 (2026-03) 35 Figure 20: Variation of the metrics used in the New KPIs planning methodology as a function of the combined transmission and reception antenna gains It is remarked that, while the example discussed in the present clause 6 involves selecting only the size of the antennas to complete the link design process, other planning scenarios could be more complex and involve candidate radio configurations with also variations in the total output powers provided at the antenna port, arising, for instance, from the availability of different product versions or the implementation of Remote Transmit Power Control (RTPC) policies - which ultimately result in limitations on the effective radiated power and are thus sensitive to the employed transmission antenna gains. Nevertheless, the approach outlined in table 6 can be easily extended to account for these additional cases. |
314690db30e0da820145d5cd1eef3bc5 | 104 141 | 6.5 Link planning with unknown traffic distribution | When the information on the link's target traffic demand distribution is not available, the BTA computation should be based on the conservative but effective approach described in clause 4.3. Table 8 illustrates all the steps needed to plan any backhaul link according to the New KPIs methodology in this case. Table 8: Link planning procedure according to the New KPIs methodology with unknown cumulative distribution function of the link's target traffic demand Initialization 1) For each th radio link configuration, compute the following metric: (24) as the sum of the maximum transmit power (in dBm), the transmit antenna gain (in dB) and the receive antenna gain (in dB) of the available radio equipment; 2) sort the available radio link configurations in ascending order based on the metric , to obtain: 1 2 β― , (25) being the total number of radio link configurations; 3) initialize 1; ETSI ETSI TR 104 141 V1.1.1 (2026-03) 36 Link Planning 4) compute the metrics prescribed by the New KPIs methodology considering the th radio link configuration: 4a) derive the PIR fade margin ( ) as: ( ) = + + ( ) β β β ( ), (26) where is the transmit power of the radio equipment relative to the PIR (in dBm), an are defined as in step 1, is the free-space path loss (in dB), is the attenuation due to atmospheric gases (in dB), and ( ) is the receiver sensitivity threshold relative to the PIR (in dBm); NOTE: Term can be computed according to Recommendation ITU-R P.676-13 [i.5]. 4b) derive the availability ( ) of each th backhaul capacity (with ( ) < ( ) < . . . < ( )) according to the well-established methodologies in Recommendation ITU-R P.530-19 [i.4]; 4c) derive the BTA lower bound by following the analytical procedure detailed in clause 4.3 (table 1), using as inputs the current availabilities ( ) (computed in step 4b) and the average and the peak values of the expected traffic demand [] and , respectively; 5) if the PIR fade margin ( ), the BTA lower bound and the CIR availability derived in step 4 meet the planning targets, then end the procedure and plan the link with the th radio configuration, otherwise increment the counter by 1; 6) If β€ repeat the process starting from step 4, otherwise end the procedure and conclude that no radio link configuration among the available ones can meet the planning targets. Table 9 reports the values of the PIR fade margin ( ), the BTA lower bound and the CIR availability that can be derived by applying steps 4a through 4c of the methodology in table 8 to each th radio link configuration. Comparing these results with the New KPIs target conditions specified in clause 6.3 clearly indicates that, also in this case, the overall planning procedure in table 8 would designate configuration 2 as the preferred radio technology for the link under study. Table 9: Key results obtained by applying steps 4a through 4c of the planning procedure illustrated in table 8 to the radio configurations described in clause 6.3 Radio link configuration PIR fade margin () BTA lower bound CIR availability = 1 9,26 dB 99,955 % 99,995 % = 2 15,28 dB 99,974 % 99,997 % = 3 21,30 dB 99,984 % 99,998 % For the sake of clearness, the results obtained from the application of the analytical method of clause 4.3 are here reported for the only case of configuration 1 (first row of table 9). Specifically, table 10 shows the set of the test parameters selected in the present example (as per step 2 of the procedure in table 1), the corresponding parameters derived from equation (12), the indication on whether constraint (13) is satisfied or not, and the resulting BTAs . In the examined scenario, the analytical method would yield a BTA lower bound equal to 99,955 %, which corresponds to the minimum value among the BTAs displayed in the fourth column of table 10 that satisfy condition (13). ETSI ETSI TR 104 141 V1.1.1 (2026-03) 37 Table 10: Results obtained from the application of the analytical method proposed in clause 4.3 to the radio link configuration 1 β from equation (12) Is constraint (13) satisfied? BTA [%] 0,05 0,079 No 99,891 0,1 0,157 No 99,902 0,2 0,315 No 99,918 0,3 0,472 No 99,929 0,4 0,629 No 99,937 0,5 0,786 No 99,942 0,6 0,944 No 99,946 0,7 1,101 No 99,949 0,8 1,258 No 99,952 0,9 1,415 No 99,953 1 1,573 Yes 99,955 2 3,145 Yes 99,961 3 4,718 Yes 99,963 4 6,290 Yes 99,963 5 7,863 Yes 99,964 10 15,726 Yes 99,965 20 31,451 Yes 99,966 40 62,902 Yes 99,967 60 94,353 Yes 99,967 62 97,498 Yes 99,967 64 100,643 Yes 99,967 66 103,788 Yes 99,967 68 106,934 No 99,967 70 110,079 No 99,967 |
314690db30e0da820145d5cd1eef3bc5 | 104 141 | 7 Conclusions | Planning backhaul networks through the New KPIs methodology calls for reliable techniques for predicting the traffic demand distributions across the different links, which are essential for accurately evaluating the novel BTA metric. The present document has introduced two complementary solutions to address this challenge. The first method relies on an analytical procedure designed to estimate the worst-case BTA of any backhaul link. Its key advantage is that it is based solely on the knowledge of the average and peak values of the expected link traffic demand, rather than the full throughput statistical distribution. Although this approach inherently yields a conservative BTA evaluation, an experimental validation activity based on traffic time series collected from operational wireless transport networks has demonstrated its effectiveness in significantly reducing link over-engineering compared to current planning criteria. The second method involves conducting measurement campaigns on live backhaul systems with the aim of deriving a dataset of traffic demand distributions to be possibly used as reference statistics for the assessment of the BTA in various deployment scenarios, with different environmental (e.g. in terms of topology and subscribers' density) and technological (e.g. in terms of transported RAN configurations) conditions. It is worth noting that these two strategies can also be employed in a synergistic manner. In some cases, the data collected in real radio fixed networks could provide valuable insights into the traffic loads which may be expected in specific contexts of interest, that can in turn be used as inputs to the analytical procedure for computing worst-case (or lower bound) BTA values. The contributions described in the present document - aimed at defining practical methodologies for assessing the BTA metric - are believed to represent a significant advancement in promoting and accelerating the adoption of the New KPIs-based design paradigm in current and future wireless backhaul networks. ETSI ETSI TR 104 141 V1.1.1 (2026-03) 38 Annex A: Experimental validation of the analytical procedure for deriving BTA lower bounds: methodology and results A.1 Overview Annex A provides a detailed description of the methodology used to define constraint (13) in the analytical procedure for deriving BTA lower bounds as illustrated in table 1. As already mentioned in clause 4.3, inequality (13) limits the search of the BTA lower bounds within the subspace of (, ) parameters that generate plausible traffic demand distributions where high throughput values occur with progressively smaller probabilities with respect to lower capacities, thus excluding, by way of example, the unrealistic behaviours previously highlighted in figures 4 and 5 (for = 0,5 in figure 4-(a) and = = 0,5 in figure 5-(a)). More specifically, this condition has been imposed by setting a restriction on the slope of the admissible cumulative distribution functions in the high-traffic region, and, in mathematical terms, it has been achieved by requiring that the maximum value of the derivative of the CDF , , produced by a generic pair of parameters (, ) and evaluated at a traffic load equal to Γ (say with 0,9 β€ β€1) is upper-bounded by a convenient term /: ,, Γ = = Γ , , = Γ () Ξ + β€ , (A.1) where , , is the probability density function associated with the pair of parameters (, ), and is given by equation (7). The choice = 0,999 and = 0,05 for constraint (13) has been determined as a result of a thorough simulation campaign based on a dataset of traffic time series acquired from a currently operative backhaul network, as further detailed below. The remainder of annex A is organized as follows. Clause A.2 provides an overview of the employed database, while clause A.3 describes the methodology adopted for the simulation activity, in terms of main assumptions, utilized metrics and test cases. Finally, clause A.4 discusses the numerical results. A.2 Database description The results described in annex A are based on a database composed of 1 510 time series obtained by recording the peak traffic loads of a set of links deployed in a commercial backhaul network for an overall period of 18 weeks, with a time resolution of 1 hour. It is remarked that this coarse data resolution (e.g. with respect to the 15 minutes time granularity typically supported by performance monitoring systems in currently deployed backhaul networks) does not constitute a limitation for the following analysis, since the real purpose of annex A is to numerically (and massively) assess how effectively the proposed analytical procedure is able to derive accurate lower bounds on the BTAs of transport links with arbitrary - yet realistic - traffic distributions, rather than to achieve a New-KPIs-compliant link planning (a process that would have undoubtedly involved managing traffic data with time resolutions on the order of 1 second, according to ETSI GR mWT 028 [i.1] and as discussed in different occasions within the present document). In order to prevent the present study from being biased in favour of traffic distributions that would not be realistically and practically considered in any rationale backhaul link design process, the time series composing the above-mentioned database have been selected so as to: β’ have a Peak-to-Average Ratio (PAR) lower than or equal to 8; β’ have analytically continuous cumulative distribution functions; β’ have cumulative distribution functions with probabilities higher than 99 % and 99,8 % for traffic values greater than 0,99 Γ and 0,998 Γ (being the maximum value of each time series), respectively, thus ensuring a smooth behaviour in the high-throughput region (figure A.1 illustrates graphically this condition). ETSI ETSI TR 104 141 V1.1.1 (2026-03) 39 NOTE: Only the time series characterized by cumulative distribution functions with probabilities higher than 99 % and 99,8 % for traffic values greater than 0,99 and 0,998 , respectively, have been included in the database. Figure A.1: Graphical representation of the condition imposed on the cumulative distribution functions of the selected time series A.3 Methodology, system assumptions and test cases Each time series in the database described in clause A.2 has been re-scaled so as to have: ,, (A.2) where parameter 1 accounts for the unavoidable gap between the actual peak traffic value and the maximum capacity , that can be delivered by the employed radio transport technology, and it will be varied in the different test cases considered in the present study. Backhaul links operating in E-band (at around 80 GHz) over a bandwidth of 500 MHz with vertically polarized electromagnetic field and supporting an ACM policy are here considered as the reference transport technology. More specifically, the analysis illustrated in clause A.4 will focus on five representative links, with hop lengths 1, 2, 3, 4 and 5 km, all deployed in a region with rainfall intensities exceeding 42 mm/h for 0,01 % of the time in a year. A total of nine test datasets have been derived from the original database described in clause A.2: Test dataset I: composed of all the traffic time series (i.e. with PAR β€ 8), and selecting 0,82 in equation (A.2); Test dataset II: composed of all the traffic time series, and selecting 0,9; Test dataset III: composed of all the traffic time series, and selecting 0,95; Test dataset IV: composed of the only traffic time series with PAR β€ 6, and selecting 0,82; Test dataset V: composed of the only traffic time series with PAR β€ 6, and selecting 0,9; Test dataset VI: composed of the only traffic time series with PAR β€ 6, and selecting 0,95; Test dataset VII: composed of the only traffic time series with PAR β€ 4, and selecting 0,82; Test dataset VIII: composed of the only traffic time series with PAR β€ 4, and selecting 0,9; ETSI ETSI TR 104 141 V1.1.1 (2026-03) 40 Test dataset IX: composed of the only traffic time series with PAR β€ 4, and selecting = 0,95. For each th test dataset ( = I, II, III, IV, V, VI, VII, VIII, IX): 1) a cumulative distribution function has been first computed for each βth traffic time series, and has been then employed to generate, by applying equation (1), the actual link BTA value for each considered hop length : β, (β= 1,2, β¦ , ; = 1, 2, 3, 4, 5 km), (A.3) where dependence on the overall antenna gain (namely, including both the receive and the transmit side), in dB, is explicitly indicated to facilitate the following description, while is the total number of time series included in the th test dataset. NOTE 1: To derive the actual link BTAs, the availabilities of the different capacities offered by the radio connection for each hop length have been computed on the basis of Recommendation ITU- R P.530-19 [i.4] by taking into account free-space loss, gaseous absorption and rain attenuation as the primary propagation impairments. 2) an average and a maximum traffic load value ([] and , respectively) have been derived for each βth traffic time series in the dataset, and have been employed as inputs to the analytical procedure as detailed in table 1, that has been run for each link distance and for each th choice of a pre-defined set of pairs of parameters ( , ) to be used in constraint (13), in order to derive the corresponding BTA lower bound: β,,( ) (β= 1,2, β¦ , ; = 1, 2, 3, 4, 5 km; = 1,2, β¦ ). (A.4) The analysis presented in clause A.4 will evaluate all the combinations of β 0,99, 0,999, 0,9999, 0,99999 and β 0,002, 0,004, 0,006, 0,008, 0,01, 0,02, 0,03, 0,04, 0,05, 0,06, 0,07, 0,08, 0,09, 0,1, 0,2, 0,22, 0,24, 0,26, 0,28, 0,3, 0,32, 0,34, 0,36, 0,38, 0,4, 0,5, 0,6, 0,7, 0,8, 0,9, 1, leading to an overall cardinality = 4 Γ 31 = 124. 3) a relative error between the derived BTA lower bound (step 2) and the actual link BTA (step 1) has been then computed for each βth time series, for each link distance and for each th pair ( , ) as: β,, = β (1 β β,( )) β(1 β β,, ) (1 β β,( )) . (A.5) NOTE 2: In formula (A.5), the relative error is expressed in terms of the outage quantities (1 β β,( )) and (1 β β,,( )). 4) an excess gain Ξβ,, in dB needed to achieve a BTA lower bound β,, numerically equal to the actual link BTA β, has been computed for each βth time series, for each link distance and for each th pair ( , ). In mathematical notation, excess gain Ξβ,, guarantees that: β,, + Ξβ,, = β, . (A.6) It is highlighted that the latter metric has the scope of quantifying, in [dB], the superfluous system margin that the employment of the proposed analytical procedure for deriving BTA lower bounds with parameters ( , ) would introduce in a possible planning process involving a link with length and carrying a traffic expressed by the βth time series. In clause A.4, the best pair of parameters ( , ) is investigated for every th test dataset on the basis of an assessment of the following three key metrics: 1) the statistical distribution of the relative errors β,, derived for all the time series and the five link distances under consideration; 2) the statistical distribution of the excess gains Ξβ,, derived for all the time series and the five link distances under consideration; ETSI ETSI TR 104 141 V1.1.1 (2026-03) 41 3) the lower bound efficiency , defined as the percentage of cases in which the proposed analytical procedure employing the th pair ( , ) in constraint (13) succeeds in generating actual BTA lower bounds. Since, according to equation (A.5), β,,( ) constitutes a lower bound on the actual link BTA value computed for each βth time series and for each link distance if and only if β,, β₯0, it yields: = 1 Γ 5 β,, Γ 100 [%], β (A.7) where () is the step function, defined as: = 1 β₯0 0 < 0 (A.8) A.4 Numerical results The numerical results obtained for Test datasets I through IX are summarized in tables A.2 through A.10, respectively, which show, for each th choice of the pair of parameters ( , ) (i.e. for each th row): i) the lower bound efficiency (third column); ii) the minimum relative error experienced over all the time series and the five link distances (fourth column); and iii) the median (fifth column), the 75th percentile (sixth column), the 90th percentile (seventh column), the 95th percentile (eighth column) and the 99th percentile (ninth column) of the statistical distribution of the excess gains Ξβ,, derived for all the time series and the five link distances (i.e. β= 1,2, β¦ , and = 1,2,3,4,5 km). It is observed that, according to equation (A.5), the minimum relative error in the fourth column is negative whenever the efficiency metric is strictly lower than 100 %, and in those cases it provides an important indication of the extent to which an actual link BTA value can fall below the BTA lower bound computed according to the proposed method with the th pair ( , ) used in constraint (13). NOTE: For conciseness, tables A.2 through A.10 only report a sub-selection of the analysed values of parameters ( , ). Tables A.2 through A.10 illustrate that employing the analytical procedure with relaxed conditions on constraint (13) (for example, in rows 3-4 of table A.2) guarantees to find actual BTA lower bounds for all the links ( = 100 %), at the expense of a generally high excess gain (larger than 5 dB in 50 % of cases in the considered examples). At the same time, the results suggest that the inherent trade-off between lower bound tightness and overall accuracy can be optimized by pursuing the reasonable compromise of selecting pairs of parameters ( , ) that could cause a slight degradation in the lower bound efficiency (with respect to the full scale of 100 %), while yielding the advantage of maintaining the excess gain below 3 dB in the majority (i.e. in the range 95 % to 99 %) of cases. It is remarked that the 3 dB value is compatible with the margin that is typically considered in network planning phases to compensate for unexpected system losses along the transmission links, such as those caused by misalignments between transmit and receive antennas due to imperfect installation. Accordingly, in the present document the optimum pair ( , ) is defined as the one that results in the minimum 95th percentile value of the excess gain across all test cases, while simultaneously guaranteeing the following two conditions: C1 lower bound efficiency always higher than 99 %; C2 minimum relative error experienced in all test cases higher than -21 % (it is recalled that the relative error is negative only when the BTA lower bound value is higher than the actual link BTA, according to equation (A.5)). It is noted that conditions C1 and C2 imply that the proposed analytical procedure can fail to compute an actual lower bound in only 1 % of the cases within the considered datasets, and that, in such rare circumstances, the BTA lower bound can exceed the actual link BTA by only a very limited amount. To illustrate this quantitatively, table A.1 reports, for each BTA lower bound value shown in the first column, the minimum actual link BTA that could result according to condition C2 and equation (A.5). ETSI ETSI TR 104 141 V1.1.1 (2026-03) 42 Table A.1: BTA lower bounds and corresponding minimum actual link BTAs according to condition C2 BTA lower bound [%] Minimum actual link BTA [%] 99,9 99,873418 99,95 99,936709 99,99 99,987342 99,995 99,993671 99,999 99,998734 Considering all the analysed test cases, the optimum pair of parameters , that minimizes the 95th percentile value of the excess gain and that at the same time guarantees conditions C1 and C2 turns out to be: 0,999, 0,05, (A.9) and this is the recommended choice for the application of the analytical procedure in table 1 of clause 4.3. Table A.2: Selected results for Test dataset I (PAR β€ 8, Ο = 0,82) ETSI ETSI TR 104 141 V1.1.1 (2026-03) 43 Table A.3: Selected results for Test dataset II (PAR β€ 8, Ο = 0,9) Table A.4: Selected results for Test dataset III (PAR β€ 8, Ο = 0,95) ETSI ETSI TR 104 141 V1.1.1 (2026-03) 44 Table A.5: Selected results for Test dataset IV (PAR β€ 6, Ο = 0,82) Table A.6: Selected results for Test dataset V (PAR β€ 6, Ο = 0,9) ETSI ETSI TR 104 141 V1.1.1 (2026-03) 45 Table A.7: Selected results for Test dataset VI (PAR β€ 6, Ο = 0,95) Table A.8: Selected results for Test dataset VII (PAR β€ 4, Ο = 0,82) ETSI ETSI TR 104 141 V1.1.1 (2026-03) 46 Table A.9: Selected results for Test dataset VIII (PAR β€ 4, Ο = 0,9) Table A.10: Selected results for Test dataset IX (PAR β€ 4, Ο = 0,95) ETSI ETSI TR 104 141 V1.1.1 (2026-03) 47 Annex B: A methodology for analysing the impacts of New KPIs on Total Cost of Ownership The aim of the present annex B is to illustrate a methodology for obtaining a general assessment of the benefits, in terms of cost of ownership, brought by the New KPIs paradigm. The envisioned approach starts by identifying a target backhaul network and a desired peak capacity that needs to be transported over each link (for the sake of simplicity - and without affecting the validity of the presented methodology - a unique peak capacity value for all the links is here considered). Based on these inputs, the cost savings enabled by the New KPIs planning methodology - with respect to traditional designs - can be evaluated according to the following step-by-step procedure: 1) derive the statistical distribution of the link lengths of the target backhaul network with a predefined granularity, e.g. obtaining a relative frequency histogram qualitatively similar to the one shown in figure B.1 (therein with 1 km granularity); NOTE: This figure illustrates the percentages of links spanning a distance that falls within each th interval (for 1, 2, β¦ , 12). Figure B.1: Qualitative representation of the relative frequency histogram of link lengths in the target backhaul network 2) identify a set of preferred transport technologies that are able to deliver the desired peak capacities; 3) define the target conditions for both the traditional and the New KPIs planning methodologies. By way of example, table B.1 shows a possible set of objectives to be guaranteed for the two approaches; 4) for each transport technology as selected in step 2, identify the maximum link distances that can be achieved by employing both the traditional and the New KPIs planning methodologies; 5) compute the Total Cost of Ownership (TCO) of the target backhaul network planned according to the traditional planning methodology as: , (B.1) where index varies across the resolvable intervals of the link lengths distribution as derived in step 1 ( 12 in the example of figure B.1), is the cost of the least expensive transport technology that can be employed to cover all the connection distances included in the ith interval while satisfying the target conditions (e.g. those outlined in the first row of table B.1), while is the relative number of links in the network with distances included in the ith interval (see figure B.1 for a graphical representation); ETSI ETSI TR 104 141 V1.1.1 (2026-03) 48 6) similarly to step 5, compute the TCO of the target network planned according to the New KPIs methodology as: Μ ; (B.2) 7) compare the TCOs derived at steps 5 and 6 to determine the cost savings enabled by the New KPIs planning methodology. NOTE: Within each th link lengths interval, the set of the eligible transport technologies depends on the maximum distance thresholds derived in step 4, therefore Μ generally differs from parameter used in step 5. Figure B.2: Link lengths distribution of the backhaul network analysed as example in the present annex B Table B.1: Target conditions for the traditional and New KPIs planning methodologies considered in the present annex B Target conditions Traditional planning β’ β₯ 99,995 % availability for a capacity representing 15 % to 20 % of the desired peak capacity β’ β₯ 5 dB fade margin for the modulation format providing the desired peak capacity New KPIs planning β’ β₯ 99,995 % availability for a capacity at least equal to 200 Mbit/s β’ β₯ 99,9 % BTA β’ β₯ 5 dB fade margin for the modulation format providing the desired peak capacity For the sake of clarity, the remaining part of the present annex B will apply the procedure described above to assess the potential savings - in terms of costs related to the yearly spectrum license fees only - enabled by the New KPIs paradigm. Towards this goal, a hypothetical backhaul network characterized by a link lengths distribution as shown in figure B.2 is considered, and two target European deployment regions are taken into account, with rainfall intensities exceeding either 32 mm/h or 42 mm/h for 0,01 % of the time in a year (these will be referred to in the following as 32 mm/h and 42 mm/h rain intensity zones, respectively, for simplicity). Pursuing a desired peak capacity of 4 Gbit/s for all the connections, three backhaul solutions have been selected as candidate technologies: β’ an E-band point-to-point system with 250 MHz bandwidth and dual polarization; β’ a Dual Band point-to-point system operating on both E-band (250 MHz bandwidth, dual polarization) and 18 GHz frequency band (56 MHz bandwidth, dual polarization); ETSI ETSI TR 104 141 V1.1.1 (2026-03) 49 β’ an 18 GHz point-to-point system aggregating two 112 MHz channels and operating with dual polarization. Tables B.2 and B.3 quantify the maximum link lengths achievable by the three different transport technologies described above for 32 mm/h and 42 mm/h rain intensity zones, respectively, considering the target conditions shown in table B.1 (it is remarked that, in the present analysis, the assessment of the BTA has been based on the analytical procedure detailed in table 1 for all the links). Based on these results, the numerical evaluations that follow have been limited to the subset of links spanning a distance lower than or equal to 10,5 km and 10 km for the 32 mm/h and 42 mm/h rain intensity zones, respectively, which represent the largest achievable hop lengths under the traditional planning methodology in the two cases. The New KPIs, enabling a wider adoption of both the E-band and the Dual Band technologies compared to traditional planning approaches, position themselves as the most favourable metrics for reducing the overall spectrum-related expenditures in scenarios where yearly license fees are inversely proportional to the operational frequency - a characteristic that applies to the majority of cases in today's backhaul markets (by way of example, table B.4 reports the per-link annual spectrum costs for the three backhaul technologies under inspection in two European Countries). In line with this consideration, the present analysis has revealed significant savings of: β’ 22 % and 26 % in the 32 mm/h rain intensity zone when considering the per-link yearly spectrum license fees of EU Country 1 and 2 in table B.4, respectively; β’ 28 % and 35 % in the 42 mm/h rain intensity zone when considering the per-link yearly spectrum license fees of EU Country 1 and 2, respectively. It is also observed that, compared to backhaul technologies operating at lower frequencies, E-band systems require fewer radio channels to achieve the same delivered capacities, thanks to their significantly larger available bandwidths. Consequently, the adoption of the New KPIs approach - which extends the E-band usability range - offers the additional benefit of a more cost-effective deployment due to the generally reduced amount of hardware to be installed. Table B.2: Maximum achievable link lengths with traditional and New KPIs planning methodologies for 32 mm/h rain intensity zone Traditional planning New KPIs planning E-band 4,2 km 5,3 km Dual Band 8,8 km 14,3 km 18 GHz 10,5 km 10,5 km Table B.3: Maximum achievable link lengths with traditional and New KPIs planning methodologies for 42 mm/h rain intensity zone Traditional planning New KPIs planning E-band 3,3 km 4,2 km Dual Band 7 km 11,5 km 18 GHz 10 km 10,5 km Table B.4: Per-link annual spectrum license fees for the three backhaul technologies under inspection in two European Countries EU Country 1 EU Country 2 E-band 575 β¬ 1 614 β¬ Dual Band 1 955 β¬ 4 924 β¬ 18 GHz 3 850 β¬ 12 841 β¬ ETSI ETSI TR 104 141 V1.1.1 (2026-03) 50 Annex C: Method for determining the target BTAs on individual links in tree-shaped backhaul network topologies C.1 Method for a simple network topology Figure C.1: Backhaul network topology analysed in the present clause C.1 The present annex C focuses on the network topology illustrated in figure C.1, where four radio sites: , , , are connected to a common aggregation node via the distinct sets of links: β 1, β 2,1, β 3,1, β 4,3,1, respectively. It is assumed that the transport network under consideration needs to be planned to ensure that the traffic generated by each radio site ( 1,2,3,4) and directed to node achieves an end-to-end (i.e. over the whole cascade of connections belonging to set β) BTA equal to . To meet these conditions, the target BTAs of the different backhaul links should satisfy the set of inequalities obtained by applying the BTA apportionment rule for daisy-chain topologies as described in annex A of ETSI GR mWT 028 [i.1] successively to all radio sites. More specifically, in the scenario of figure C.1: β’ radio site is connected to the aggregation node exclusively via link 1. Consequently, to meet the constraint on the objective end-to-end BTA , link 1 should be designed to achieve a target Backhaul Traffic Availability such that: 1 1 ; (C.1) β’ radio site reaches node through links 2 and 1. Therefore, the target BTAs for link 2 and link 1, denoted as and , respectively, should satisfy the following inequality: 1 1 1 ; (C.2) β’ radio site is connected to node via links 3 and 1. The corresponding target BTAs - and - should then satisfy: 1 1 1 ; (C.3) β’ radio site communicates with node through links 4, 3, and 1. The target BTAs for links 4, 3 and 1 - namely , and - should satisfy: 1 1 1 1 . (C.4) ETSI ETSI TR 104 141 V1.1.1 (2026-03) 51 Any selection of the target link BTAs - , , and - that satisfies the constraints in inequalities (C.1) through (C.4) ensures a valid network design capable of meeting the desired end-to-end Backhaul Traffic Availability levels , , and for the data flows generated by each node. When a uniform BTA apportionment is applied across the sequence of links connecting each radio site to the aggregation node , the inequalities (C.1) through (C.4) result in the following constraints on the individual target link BTAs : β’ for (single link): 1 β β€ 1 β (C.5) β’ for (two links): 1 β β€ 1 β 2 , 1 β β€ 1 β 2 (C.6) β’ for (two links): 1 β β€ 1 β 2 , 1 β β€ 1 β 2 (C.7) β’ for (three links): 1 β β€ 1 β 3 , 1 β β€ 1 β 3 , 1 β β€ 1 β 3 . (C.8) Based on constraints (C.5) through (C.8), the individual target link BTAs can be selected as follows: β’ for link 1, which is shared across all four radio sites, the most stringent constraint applies. Therefore: 1 β = 1 β , 1 β 2 , 1 β 2 , 1 β 3 (C.9) β’ for link 2, used only by radio site : 1 β = 1 β 2 (C.10) β’ for link 3, shared by sites and : 1 β = 1 β 2 , 1 β 3 (C.11) β’ for link 4, used exclusively by site : 1 β = 1 β 3 . (C.12) By applying equations (C.9) through (C.12), the network designer can determine the appropriate target BTAs for each individual link to ensure that the overall end-to-end Backhaul Traffic Availability conditions , , and are met for all the radio sites. Figure C.2 illustrates the resulting individual target BTAs for the case = = = = 99,9 %. ETSI ETSI TR 104 141 V1.1.1 (2026-03) 52 Figure C.2: Illustrative target BTAs for the individual links of the network topology in figure C.1 when = = = = 99,9 % and with uniform BTA apportionment strategy C.2 Method for all network topologies The method outlined in clause C.1 can be readily extended to an arbitrary transport network topology where a set of radio sites , , , β¦ , are connected to a common aggregation point via cascades of multiple links, as illustrated in figure C.3. In this generalized scenario, a set of inequalities analogous to those derived for the four-site topology in figure C.1 can be obtained by systematically applying the BTA apportionment rule to every individual path from each radio site to the aggregation node , regardless of the number or arrangement of intermediate links. Figure C.3: Generalized transport network topology ETSI ETSI TR 104 141 V1.1.1 (2026-03) 53 History Version Date Status V1.1.1 March 2026 Publication |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.