Overcoming the Challenges of WiMAX Deployment in 700MHz Band 1. Abstract The economics of the wireless systems is the major consideration when choosing the technology, the deployment strategy and the modes of operation. It determines the investment required to deploy a network, the expense required to operate it and the operator’s profit margin. In the context of WiMAX technology, a careful evaluation of the economics of the network is even more crucial. Since WiMAX technology was born into a tough world where competing alternatives such as ADSL, fiber and high speed cellular technologies already exist, WiMAX needs to have a significant advantage over these incumbent technologies in order to justify the initial deployment cost, the operation cost and to successfully compete over time in a reality of ARPU erosion. In order to substantiate the WiMAX business case, efficient ways for maximizing spectral efficiency are required. Two main techniques are proposed by the WiMAX forum to enhance spectral efficiency: Aggressive frequency reuse schemes and the use of multiple antenna techniques – MIMO (Multiple Input Multiple Output) and AAS (Adaptive Antenna Systems). WiMAX in 700MHz has some unique characteristics that make the deployment of a WiMAX network at this frequency band even more challenging. The main challenges are the high capacity required by the large cells characterizing the 700MHz deployment, the large size antenna arrays that might be required if MIMO and AAS are chosen as the capacity enhancing techniques and on top of all that the small spectrum slices assigned to operators in this band. Experienced wireless operators, not without reason, express their concern about whether aggressive reuse schemes will prove themselves in “Real World” deployments and what impact might the co-channel interference have on overall system capacity and QoS. In addition, there is a reasonable concern that MIMO will not be effective in many 700MHz links where LOS (Line of Sight) or NLOS (Near Line of Sight) exists and both MIMO and AAS antennas will not be a realistic option in 700MHz deployment due to the size of the antenna structure. The Cross Sector Interference Cancellation (XSIC) capacity enhancement technology, developed by Pallasium, addresses these concerns and provides a significant capacity boost to WiMAX networks, using small size conventional sector antennas. It enables aggressive reuse schemes and eliminates the dependency on MIMO for capacity enhancement. However, XSIC can also operate over MIMO, allowing the benefit of both worlds. XSIC software improves network economics and user experiences through interference cancellation and gains in spectral efficiency, client data rates, and overall system capacity. ` Overcoming the Challenges of WiMAX Deployment in 700MHz Band
Page 1 of 15
2. The Challenges of 700 MHz WiMAX Deployment WiMAX in the 700MHz band is mainly used for providing fixed and nomadic services in suburban and rural areas. In North America the 700MHz band may also be used for public safety application, including mobile subscribers. Due to the propagation characteristics of signals in 700MHz, a given area can be covered with large cells, resulting in potentially fewer base stations. However, since each cell serves a large number of subscribers, capacity soon becomes the limiting factor. In many 700MHz systems where relatively small slices of spectrum are assigned to the operator, capacity becomes a bottleneck even sooner. In order to exploit the long range propagation of the 700MHz signal and enable large size cells incorporating spectral efficiency enhancers becomes crucial. Unfortunately, the MIMO - WiMAX leading spectral efficiency enhancer, exhibits poor performance in 700MHz WiMAX due to the high LOS (Line of Sight) or NLOS (Near Line of Sight), typical to such deployments. On top of that, the antenna arrays required for MIMO (Multiple Input Multiple Output) and AAS (Adaptive Antenna Systems) in the 700MHz band are, in many situations, too large to be a realistic option.
Reuse Strategies Following the WiMAX Forum, the nomenclature for describing the frequency reuse pattern in this paper is (c, n, s); where c is the number of base station sites in a cluster, n is the number of unique frequency channels required and s is the number of sectors per base station site. Cell and Sector Reuse 1 (1,1,1)
Source: WiMAX Forum
Figure 1 : Frequency Reuse of 1 with 3-Sector Base (1,1,3) According to this strategy all available sub-channels are assigned to all sectors and cells with no spatial separation. The concept assumes that some amount of co-channel interference is tolerable as long as the interference is equally spread between the subscribers. To equally share the interference between the subscribers and avoid a situation where a single subscriber is severely affected, WiMAX, in FUSC and PUSC ` Overcoming the Challenges of WiMAX Deployment in 700MHz Band
Page 2 of 15
modes, assigns sub-carriers to sub-channels randomly. Although this method works quite well in an unloaded network, it looses its effectiveness as the network load increases. As more sub-carriers are simultaneously in use, more co channel collisions occur; resulting is both throughput degradation and lower probability for adequate SINR, a fact that hampers significantly the ability of providing adequate QoS. Cell Reuse 1 / Sector Reuse 3 (1,3,3) According to this strategy the available spectrum is divided between the three sectors in a way that each sector uses one third of the available spectrum as shown in Figure 2. Sector reuse 3 eliminates co-channel interference at the sector boundaries and significantly decreases co-channel interference between neighboring cells due to the increased spatial separation for channels operating at the same frequency. Neglecting for a moment the co-channel interference, sector reuse 1 has the potential of providing three times higher capacity than sector reuse 3. This is due to the fact that the same spectrum is reused three times in the same cell. However, in the real world, a significant overlap exists between sectors and between cells, resulting in co-channel interference in the overlapping areas. If a way will be found to cancel the co-channel interference in the overlapping areas, a significant capacity boost may be achieved.
Source: WiMAX Forum
Figure 2 : Cell reuse 1 / sector reuse 3 (1,3,3)
3. XSIC – Capacity Enhancement with No Additional Antennas and Radios The Cross Sector Interference Cancellation (XSIC) technology, developed by Pallasium, addresses these concerns and provides a significant capacity boost to WiMAX networks, using small size conventional sector antennas. It enables ` Overcoming the Challenges of WiMAX Deployment in 700MHz Band
Page 3 of 15
aggressive reuse schemes and eliminates the dependency on MIMO for capacity enhancement. XSIC can also operate over MIMO, allowing the benefit of both worlds.
How Does XSIC Work ?
Figure 3: Cross sector interference in a conventional sector
Figure 4: Cross sector cancellation when XSIC is employed
XSIC provides the operator with an efficient capacity enhancement tool using simple sector antennas. Capacity gain is achieved by allowing the operator to reuse the same spectrum by all sectors (sector reuse -1) while the XSIC algorithm eliminates the cochannel interference in the overlapping areas between the sectors. In contrast with AAS and MIMO, which become highly questionable in 700MHZ deployment, XSIC does not require additional antennas and transceivers. Its algorithm utilizes only the existing conventional sector antennas. Figure 3 illustrate cross sector interference in a three-sector cell. The desired signal is represented by a solid line and the interfering signal is represented by a dashed line. The signal originating in sector-1 is colored blue and the signal originating in sector-2 is colored green. Figure 4 shows the same scenario with the XSIC algorithm activated. The cross sector interference is canceled by steering a null toward the interferer in the pertinent pattern, so that each subscriber receives only the signal from the sector it is affiliated to. As opposed to the AAS and MIMO methods, keeping large distances between the sector antennas is not required. On the contrary, it is better for the sector antennas to be with as little spacing as possible.
Using XSIC in a Single Cell Network To avoid cross sector interference in conventional single cell deployment, sector reuse is employed. In a 3-sectors cell, for instance, each sector uses a third of the available sub-carriers. When employing XSIC algorithm in such a cell, frequency reuse-1 plan ` Overcoming the Challenges of WiMAX Deployment in 700MHz Band
Page 4 of 15
can be implemented where all three sectors use all sub-carriers. By utilizing the same spectrum three times at the same cell capacity a gain of about 300% can achieved.
1/3 1/3 spectrum spectrum
Full Full spectrum spectrum
1/3 spectrum
Full spectrum
With XSIC, the same spectrum can be reused in all sector. Resulting in close to 3 times capacity improvement.
With no XSIC, sector reuse 3 is required to avoid cross sector interference
Figure 5: Using XSIC in a single cell 3-sector system.
XSIC Capacity Improvement in a Single Cell System To demonstrate the improvement achieved by Pallasium's XSIC technology in a single cell deployment the following reuse schemes and scenarios were investigated: •
Sector reuse-3 (each sector uses one third of the spectrum without XSIC)
•
Full reuse-1 (the three sectors use the same spectrum without XSIC)
•
Full reuse-1 with XSIC (the three sectors use the same spectrum with XSIC employed).
Full details on the scenarios' parameters and simulations conditions are provided in Appendix A. The following table summaries the simulation results for single cell scenarios, in terms of spectral efficiency. It compares the spectral efficiency of sector reuse 1 and sector reuse 3 plans without XSIC to a reuse 1 plan while XSIC is activated. Spectral efficiency is provided per two criteria: calculated for both criteria: “Equal Time” and “Equal Data”. (According to “Equal Time” criteria, throughput is calculated where each subscriber is granted the same activity time. According to “Equal Data” criteria, throughput is calculated where each subscriber is granted the same data volume). Table 1: Single Cell Simulation results comparison Equal Time
Equal Data
Bit/Cell/Hz Improvement
Bit/Cell/Hz Improvement
System
Baseline Reuse1
1.9
-
0.96
-
Baseline Sector Reuse-3
1.46
-23%
1.29
34%
XSIC Reuse-1
4.34
128%
3.94
310%
` Overcoming the Challenges of WiMAX Deployment in 700MHz Band
Page 5 of 15
As expected, the fact that XSIC enables using three times the same spectrum, provides a significant spectral efficiency improvement over reuse-1 and reuse-3 scenarios: 128% for “Equal Time” criteria and 310% for “Equal Data” criteria.
Using XSIC in a Multi Cell Network The prevalent reuse strategy in a multi-cell WiMAX network is Cell Reuse 1 / Sector Reuse 3, shown in Figure 6-a. Using XSIC, a novel topology can be created which significantly improves spectral efficiency. The new topology sets the frequency reuse in cells structure rather than in sectors. In this topology cells are arranged in clusters of three. Each cell uses one third of the available spectrum. On the other hand, each of the three sectors of the cell uses the full spectrum assigned to this cell. Due to the use of XSIC technology inter-sector interference are avoided. Figures 6-a and 6-b illustrate both cell reuse 3 and sector reuse 3 approaches. As can be seen in figure 6-a and 6-b, the reuse distance in the cell frequency reuse-3 is higher by approximately 75% than sector frequency reuse-3, resulting in a corresponding spectral efficiency improvement of more than 160 %.
Figure 6-a: Conventional approach: Cell reuse 1 / Sector reuse 3. (1,3,3)
.
Figure 6-b: Better reuse scheme: using XSIC: Cell reuse 3 / Sector reuse 1, (3,3,1)
Figure 6: Using XSIC in a multi cell network
Throughput Enhancement Using XSIC Adaptive modulation assigns each subscriber the highest possible modulation scheme allowed by the quality of the link. In the presence of co-channel interference, the quality of the link degrades and a lower modulation scheme is assigned to the subscriber. The following figures show the size and location of the areas where each modulation scheme is provided.
` Overcoming the Challenges of WiMAX Deployment in 700MHz Band
Page 6 of 15
Figures 7 and 8 show the offered throughput (in terms of modulation scheme) in reuse (1,3,3) vs. reuse (3,3,1) with XSIC. As can be seen, larger areas can be served with higher modulation schemes.
Figure 7: Offered modulation schemes with “conventional” reuse (1,3,3)
Figure 8: Offered modulation schemes with reuse (3,1,3) with XSIC
` Overcoming the Challenges of WiMAX Deployment in 700MHz Band
Page 7 of 15
The results shown in Figure 7 and 8 are summarized in the following table.
Table 2: Offered Modulation Scheme with and without XSIC Under
QPSK 1/2 to
QAM16 1/2
QPSK 1/2
QAM16 1/2
QAM 64 2/3
QAM(?) 2/3 and above
Cell reuse 1 / sector reuse 3 – NO XSIC
23%
35%
20%
22%
Cell reuse 1 / sector reuse 3 – with XSIC
9%
19%
36%
36%
Reuse Plan
Spectral Efficiency Enhancement Using XSIC The following table summaries the performances of the different reuse schemes and relative performance improvement compared to the ‘Baseline reuse-1’ system. Spectral efficiency is provided per “Equal Time” and “Equal Data” criteria.
Table 3: XSIC spectral efficiency improvement in a (3,1,3) reuse scheme Reuse Scheme
“Conventional” (1,3,3) reuse scheme XSIC employed (3,1,3) reuse scheme
on
Equal Time
Equal Data
Bit/Cell/Hz Improvement
Bit/Cell/Hz Improvement
0.68
-21%
0.31
0%
1.02
18.6%
0.52
67.7%
As can be seen in Table 3 significant improvement is achieved by use of XSIC technology over conventional reuse 1 and reuse 3 plans.
Employing XSIC on a full reuse 1 (1,1,3) scheme An aggressive full reuse 1 (1,1,3) scheme is promoted by several vendors. According to this strategy, all available sub-channels are assigned to all sectors and cells with no spatial separation. This concept assumes that some co-channel interference is tolerable as long as the interference is equally spread between the subscribers and not concentrated in one or in a few of them. This method has the highest potential for “sub-carries collision” resulting in co-channel interference. When XSIX is employed on a full reuse 1 scheme, a significant improvement in terms of SINR, spectral efficiency and capacity is gained. Table 2 summarizes the simulation results in terms of percentage of subscribers for each range of modulation scheme. ` Overcoming the Challenges of WiMAX Deployment in 700MHz Band
Page 8 of 15
Figure 9: Offered modulation schemes in reuse (1,1,3) without XSIC (CPE Omni)
Figure10: Offered modulation schemes in reuse (1,1,3) with XSIC (CPE Omni)
Figures 9 and 10 show the offered throughput (in terms of modulation scheme) in reuse (1,1,3) vs. reuse (3,1,3) with XSIC. Comparing Figure 9 to Figure 10 it is easy to see that with XSIC employed on reuse (3,3,1), larger areas can be served with higher modulation schemes. The results shown in Figure 9 and 10 are summarized in the following table: Table 4: Offered Modulation Scheme with and without XSIC Under
QPSK 1/2 to
QAM16 1/2
QPSK 1/2
QAM16 1/2
QAM 64 2/3
QAM(?) 2/3 and above
Cell reuse 1 / sector reuse 1 – No XSIC
63%
24%
11.5%
1.5%
Cell reuse 1 / sector reuse 1 – with XSIC
52%
23%
14%
11%
Reuse Plan
XSIC In Full Reuse Scheme In a Fixed System. Many 700MHZ systems serve fixed subscribers with directional antennas. In such systems, the inter-cell interference level is significantly reduced and the benefits of XSIC technology are even more prominent as shown in the following data. Adaptive modulation assigns each subscriber the highest possible modulation scheme allowed by the quality of the link. In a presence of co-channel interference, the quality of the link degrades and a lower modulation scheme is assigned to the subscriber. The following diagrams show the size and location of the areas where each modulation scheme is provided. ` Overcoming the Challenges of WiMAX Deployment in 700MHz Band
Page 9 of 15
Figure 11: Offered modulation schemes with reuse (1,1,3) without XSIC (CPE directional).
Figure 12: Offered modulation schemes with reuse (1,1,3) with XSIC (CPE directional).
The results shown in Figure 11 and 12 are summarized in the following table: Table 5: Offered Modulation in Reuse1 with fix directional SUs antennas System
QAM 2/3 and above
Under
QPSK 1/2 to
QAM16 1/2
QPSK 1/2
QAM16 1/2
QAM 64 2/3
Baseline Reuse1
34%
36%
25.5%
4.5%
XSIC Reuse1
5%
15%
36%
44%
Figures 11 and 12 show the offered throughput (in terms of modulation scheme) in reuse (1,1,3) vs. reuse (3,1,3) with XSIC. Comparing Figure 11 to Figure 12 it is easy to see that with XSIC employed on reuse (1,1,3), larger areas can be served with higher modulation schemes.
The following table provides the spectral efficiency improvement when XSIC is employed on a system serving fixed subscribers. Spectral efficiency is provided per “Equal Time” and “Equal Time” criteria: ` Overcoming the Challenges of WiMAX Deployment in 700MHz Band
Page 10 of 15
Table 6: XSIC spectral efficiency improvement in a (1,1,3) reuse scheme when CPE’s use directional antennas System
Equal Time
Equal Data
Bit/Cell/Hz
Improvement Bit/Cell/Hz Improvement
Baseline Reuse1
1.71
-
0.87
-
XSIC Reuse1
3.52
106%
2.31
165%
4. Summary The XSIC technology presented in this document, addresses the challenges involved with WiMAX deployment in the 700MHZ band. It provides the operator an efficient capacity enhancement tool using simple sector antennas. Capacity gain is achieved by allowing the operator to reuse the same spectrum by all sectors (sector reuse -1) while XSIC algorithm eliminates the co-channel interference in the overlapping areas. As opposed to AAS and MIMO, which become highly questionable in 700MHZ deployment, XSIC does not require additional antennas and transceivers. Its algorithm utilizes only the existing conventional sector antennas. XSIC principle of operation was explained and data was provided showing a significant capacity enhancement, in various reuse plans, translated to lower cost of Mbps per square area, fewer base stations per network and lower operation expenditure.
` Overcoming the Challenges of WiMAX Deployment in 700MHz Band
Page 11 of 15
Appendix A –Simulation Assumptions The performance improvement achieved by XSIC technology in 700MHZ deployment scenarios is demonstrated by means of simulation. The simulations were conducted on two network types: single cell and multi cell topologies.
Single Cell Scenario In this scenario a single 3 sector 10KM radius macrocells has been assumed. With this scenario no inter-cell interference exist. Three baseline tests have been defined: •
Sector Frequency reuses-1 without XSIC.
•
Sector frequency reuse-3 without XSIC.
•
Single omni sector.
One scenario with XSIC was tested: •
Frequency reuse-1 with XSIC.
Multi Cell Scenario The simulation was performed over 19 hexagonal 3 sector 10KM radius macro-cells, where interference is evaluated in the center cell. 2-Baseline scenarios without XSIC have been defined: •
Frequency reuse-1.
•
Sector frequency reuse-3.
Three topologies with XSIC were tested against base line: •
Cell/Sector frequency reuse-1 with XSIC.
•
Sector frequency reuse – 1-Cell frequency reuse-3 with XSIC.
•
Fractional cell reuse-3 with XSIC – a frequency reuse saving scheme recommended by the WiMAX forum.
Simulation Conditions Network Topology Basic 3 sectors 19 hexagonal cells canonic network topology was employed. Table 1: Network Topology Parameters Number of cells
19
Sectors per Cell
3
Sectors boresight angles Cell Radius
(300,1500,-900) or (00,1200,-1200) 10000m
` Overcoming the Challenges of WiMAX Deployment in 700MHz Band
Page 12 of 15
Equipment Model The equipment model assumes all BS/MSs have identical characteristics. BS is modeled with a single 900 antenna per sector. Table 2: BS Model Tx Power
36dBm
Antenna Boresight Gain
16dBi
Antenna 3dB Beamwidth1
900/1200
Front to Back power ratio
25dB
Noise Figure
4 dB
Number of Antennas per Sector
1
Note: 1) 900 antennas are used in network simulation to reduce cross cell interferences when using sector reuse-3. 1200 antennas are used in a single cell deployment to increase coverage at sectors fringes. MSs are modeled with a single omni-directional antenna and 24 dBm TX power. Table 3: MS Model Tx Power
24dBm
Noise Figure
6 dB
Number of Antennas
1
` Overcoming the Challenges of WiMAX Deployment in 700MHz Band
Page 13 of 15
Channel Model Table 4: Channel Model Pathloss
Single slop model (COST 231 HATA) PL = 34.5 + 34LOG10(d) Where d is distance in meters. Lognormal distributed random variable with zero mean and 8 dB standard deviation.
Shadowing
Implementation Loss
•
Shadowing of a given user at sectors of the same cell is 100% correlated.
•
Shadowing of a given user at sectors of different cells is 50% correlated.
1 dB
WiMAX PHY Parameters The simulation is performed with the following WiMAX Air Interface parameters. Table 5: WiMAX Channel Parameters Channel Bandwidth
10 MHz
Frame Duration
5 ms
Sub-Channelization mode
PUSC
DL/UL Data Symbols1
DL:28 UL:12
Sounding symbols per Frame
2
DL/UL Data Symbols with Sounding2
DL:26 UL:12
Notes: 1) Only data carrying symbols without preamble FCH and mapping were used. 2) When Sounding is used the UL/DL symbols allocation is changed so only the DL is influenced. Adaptive Modulation The assumption is that adaptive modulation is applied according to the MS SINR, with modulation and coding available from QPSK 1/12 to QAM64 3/4. In addition HARQ is employed with up to 3 repetitions.
` Overcoming the Challenges of WiMAX Deployment in 700MHz Band
Page 14 of 15
Table 6: Modulation SNR and capacity Data rate Per Sub- Implementation Loss channel1
Modulation Required SNR [dB]
Without With Sounding Sounding
Bits/QAM Symbol QPSK 1/18 HARQ3
-9.11
.055
7466.6
6933.3
QPSK 1/12 HARQ2
-7.34
.083
11200
10400
QPSK 1/12
-4.34
.166
22400
20800
QPSK 1/8
-2.57
.25
33600
31200
QPSK 1/4
0.43
.5
67200
62400
QPSK 1/2
3.43
1
134400
124800
QPSK 3/4
6.09
1.5
201600
187200
QAM16 1/2
9
2
268800
249600
QAM16 3/4
11.58
3
403200
374400
QAM64 2/3
16.15
4
537600
499200
QAM 64 3/4
17.2
4.5
604800
561600
1 dB
Note: 1) Data rate is calculated per sub-channel per second according the number of data carrying OFDM symbols in the DL sub-frame according to Table 5. XSIC Performance Assumption The XSIC algorithm reduces interferences between different sectors of the same cell by 15dB and up 20 maximum of -25dB.
` Overcoming the Challenges of WiMAX Deployment in 700MHz Band
Page 15 of 15