1© 2017 Viavi Solutions Inc.www.viavisolutions.com
Changing the game GEOoptimize
Viavi Confidential and Proprietary Information
2© 2016 Viavi Solutions Inc.www.viavisolutions.com2
© 2020 VIAVI Confidential & Proprietary Information
GEOoptimize - Introduction
GEOoptimize brings the power of subscriber-centric, geo-located intelligence to automated network performance optimization. GEOoptimize automated predictive design can be implemented in the network to improve coverage and capacity resulting in improved QoE of the subscribers.
With over 70% of a network’s energy consumed in the RAN, GEOoptimize can also be employed to produce a design that can reduce energy consumption at cell sites, delivering significant cost savings.
Ø Automated, simplified processØ Actuates resultsØ Subscriber-centric approachØ Location Intelligence with building-level
accuracyØ Analyzes and correlates KPIs in parallelØ Predictive analytics with intelligent
searchØ Built on over a decade of experience in
Optimization
GEOoptimize features GEOoptimize benefits
Ø Directly maps to revenue and customer QoEØ Fine-tune decisions delivering improved
coverage, capacity, OpEx savingsØ Targeted optimization of servicesØ Manages complexity for faster, more
significant gainsØ Proven business results on 3G and 4G
networks
3© 2016 Viavi Solutions Inc.www.viavisolutions.com3
© 2020 VIAVI Confidential & Proprietary Information
GEOoptimize - Introduction
GEOoptimize typical use cases
Ø Improve LTE and UMTS (Data and Voice services)• Data throughput improvement• Signal strength/quality optimisation• VoLTE QoS improvement• Drops/Blocks reduction
Ø Reduce 4G data users on 3G while maintaining 3G service
Ø Carrier Re-farming
Ø Un-even distribution of the most loaded cells and least loaded cells
Ø Switch off cells whilst maintaining the network accessibility and retainability
4© 2016 Viavi Solutions Inc.www.viavisolutions.com4
© 2020 VIAVI Confidential & Proprietary Information
GEOoptimize - Parameters
The following physical network parameters can be modified:• Power
– CPICH (UMTS)– CrsGain (LTE)– Part of Sector Power (LTE)
• Antenna Electrical Tilt• Antenna Mechanical Tilt• Antenna Azimuth• Antenna Height
Customer can focus on a select set of remote parameter changes formost cost-effective improvements, e.g. only Power and E-Tilts
Predictive analytics with
intelligent search
Parameter ranges
Delivering Subscriber-centric, Location-aware, Automated Optimization
Use caseGEOoptimize Energy
6© 2016 Viavi Solutions Inc.www.viavisolutions.com6
© 2020 VIAVI Confidential & Proprietary Information
GEOoptimize Energy - Motivation
• Mobile operators are looking at ways to reduce their OpEx cost but would like to maintain the QoE fortheir subscribers. For example:
Ø Reduce the number of 3G carriers while maintaining the subscriber and network QOS until the time when the 3Gtechnology is completely switched off
Ø Due to recent COVID-19 pandemic lockdown situation, the traditional mobile traffic hotspots have moved fromcity centers, business and retail parks towards residential areas where most of the people are confined to theirhomes. With these sudden shifts in subscriber patterns, the Mobile operators are facing challenges to provide QOSwhere their subscriber concentration and usage have shifted while maintaining service quality in the traditionalmobile traffic hotspots, especially for emergency services.
• Mobile operators are therefore looking for an optimal RAN design for their network, where they canidentify radio cells that can be switched off to achieve cost savings while maintaining subscriberservices quality of experience.
• The VIAVI automated GEOoptimize solution addresses these challenges and has proven OpEx savingresults while maintaining network QOS for a large mobile operator in Europe.
7© 2016 Viavi Solutions Inc.www.viavisolutions.com7
© 2020 VIAVI Confidential & Proprietary Information
GEOoptimize Energy – How it works
Ø GEOoptimize simulation platform uses Layer transition capability to supportmodelling for the following Energy saving use cases:§ Permanent Cell switch off
§ Permanent Site switch off§ Periodic Cell/Site switch off / switch on§ Refarming (supporting both Energy plus Data/Voice use cases)
Ø Layer transition select alternate layers/technologies which have been reported inother measurement reports in the same bin.
Ø Builds distributions of signal strength/quality for all other cells visible in the bin.Ø Reselects to alternate layers using a configurable percentile on the distribution and
a hysteresis.Ø Energy saving goal is to switch off cells whilst maintaining the network accessibility
and retainability KPIs
GEO
opti
miz
e U
I
8© 2016 Viavi Solutions Inc.www.viavisolutions.com8
© 2020 VIAVI Confidential & Proprietary Information
GEOoptimize Energy Vertical Layer Transitions – How it works
Before simulation: GEOoptimize records the UEreal measurements and locations. The UE in thereport is served by a cell at site 1 in carrier f1(Green colour).
During simulation: changes to the measuredcells’ power and quality are modelled. At thesame time, distributions of all the cellsmeasured by other users in the same bin arebuilt – covering other carriers and othertechnologies.
After simulation: users call segment with newmeasurement report is now served by preferredcell at site 1 – in a different carrier/technology(Purple colour).
Users can be served by other preferred carriers /sites and hence some cells or complete sites canbe turned off permanently or temporarily.
site 1, f1
site 1, f2
9© 2016 Viavi Solutions Inc.www.viavisolutions.com9
© 2020 VIAVI Confidential & Proprietary Information
GEOoptimize Energy Horizontal Layer Transitions – How it works
Before simulation: GEOoptimize records the UEreal measurements and locations. The UE in thereport is served by a cell at site 1 in carrier f1.
During simulation: During simulation changes tothe measured cells power and quality aremodelled and new cell at site 2 has improvedquality.
After simulation: users call segment withupdated measurement report is now served bypreferred cell at site 2 - still in carrier f1.
This type of transition uses information onlyfrom the original call segment and itsmeasurement report.
site 2, f1
site 1, f1
10© 2016 Viavi Solutions Inc.www.viavisolutions.com10
© 2020 VIAVI Confidential & Proprietary Information
GEOoptimize Energy – Use case example
Ø In the next slide VIAVI has conducted a real network OpEx saving POC to switch off cells in a cluster. The mobile operator goal was to: • Reduce OpEx• Improve energy consumption• Simplify Network• Improve or maintain service quality
Ø The same principle can be applied for OpEx savings during this COVID-19 pandemic lockdown situation. The traditional hotspots have less traffic and during the lockdown Mobile operators can apply a GEOoptimize design to switch off carriers or complete sites; while maintaining the QoE of their subscribers.
RNCs selected to collect data for the simulation
11© 2016 Viavi Solutions Inc.www.viavisolutions.com11
© 2020 VIAVI Confidential & Proprietary Information
GEOoptimize Energy – Use case example
Customer Goal
• Reduce the number of 3G cells
• Maintain KPI level for blocks and drops
Cluster information• Tech: UMTS• Sites: 42• Cells: 328
• RNCs covering optimization area: 3
Design Constraints:• CPICH limited between 30dB and 34dB• Electrical tilt restrict to +/- 3 degrees from the current position
• No tilt changes for antenna sharing GSM band
Cluster selected for GEOoptimize Energy saving design
12© 2016 Viavi Solutions Inc.www.viavisolutions.com12
© 2020 VIAVI Confidential & Proprietary Information
GEOoptimize Energy – Use case example
GEOoptimize design consideration:
• Allow optimization algorithm to switch off cells and sites
• Adjust CPICH Power and E-tilt of neighbouring cells to compensate for the effect of the site switch-offs
• Set degradation limits on blocks and drops• Approximately 30% of the cells within the cluster
have been changed by this proposed design.
Increase Up Off
CPICH 82
E-Tilt 4
Cells switch off 8
0
10
20
30
40
50
1dB 2 dB 3 dB
Change Distribution: CPICH increase
GEOoptimize Design recommendations
13© 2016 Viavi Solutions Inc.www.viavisolutions.com13
© 2020 VIAVI Confidential & Proprietary Information
GEOoptimize Energy – Use case results
Design implementation resultsØ OpEx saving result
§ Turned off 8 cells (2.5% of the cluster)
Ø No degradation in network KPIs
Ø The design has been implemented in the live networkand these cells have been permanently switched off.
Predictions
0
5000000
10000000
15000000
0 10 20 30
EcNo Histogram
After Before
• Reduce OpEx• Improve energy consumption• Simplify Network• Improve or maintain service quality
0500000
100000015000002000000
>= -4
0-4
9 - -
48-5
8 - -
57-6
7 - -
66-7
6 - -
75-8
5 - -
84-9
4 - -
93-1
03 -
-102
-112
- -1
11
RSCP Histogram
After Before
© British Telecommunications plc 2019
EE’s use of GEOoptimize
© British Telecommunications plc 2018
EE’s objectiveTo evaluate GEOoptimize’s ability to bring step-change improvement in critical KPIs and customer QOE in an efficient and economical manner.
Pilot objectiveTo improve LTE network performance in two clusters which had experienced unsatisfactory NPS scores.– Reduce VoLTE drops and blocks, and increase Average DL Throughput in the two clusters.– In this presentation the two clusters are referred to as A and B.
VIAVI’s Pilot proposal:Viavi proposed its unique GEOoptimize solution, which harnesses the power of subscriber-centric geolocated data from the existing NITRO GEO platform to drive its optimization algorithms.
Post-Pilot activity:The pilot were implemented in the EE network and the results were impressive enough for EE to select GEOoptimizefor wider deployment.EE is now working through the network cluster by cluster, based on a prioritisation of key areas (by NPS, known issues, complaints).
GEOoptimize Pilot at EE
© British Telecommunications plc 2019
Cluster A
16
© British Telecommunications plc 2018
Cluster A Executive Summary
© 2018 VIAVI Confidential & Proprietary Information 17
The goal was to improve :• Spectral efficiency (Avg. DL Throughput)• VoLTE Blocks and drops• RSRQ
Cluster information• LTE Sites: 46• LTE Cells: 404• LTE Band 1800/2600/800
Design Constraints:• RS power increase allowed for 2600 band (power decrease not allowed)• RS power increase & decrease allowed on 1800 and 800 bands; hardware and EIRP limits were considered in
recommendations• Electrical tilt changes:
- 1800/800 bands only (-2/+4 deg) up-tilt and down-tilt allowed- 2600 band only (+4 deg) down-tilt allowed, no up-tilt allowed
• Cell edge RXLEV/RXQUAL coverage targets for GSM1800 not provided
Cluster A:Mixed residential andbusiness area in London with major rail line running east-west at the north end through challenging topology, including 2 major rail stations.
© British Telecommunications plc 2018
GEOoptimize LTE Design Prediction Cluster A
© 2018 VIAVI Confidential & Proprietary Information 18
GEOoptimize predicted that by actuating its tilt and power recommendations:
• Average Spectral Efficiency would increase by 1.6 %
• Average RSRQ would increase by 0.81 %
• The estimated absolute number of drops for data sessions would reduce by 1.95 %
• QC1 estimated absolute number of drops would decrease by 4.86%
• QC1 Spectral Efficiency would improve by 2.89 %
This table shows predictions made by GEOoptimize on the impact of
the design implementation
Submitted design on 29th October
KPI Start Value
Predicted Value
Delta Percentage
Average RSRP -87.47 -88.81 -1.53
Average RSRQ -8.61 -8.54 0.81
Average Spectral Efficiency 1.25 1.27 1.6
Estimated Drops 58.18 57.04 -1.95
QCI 1 Average RSRP -90.75 -91.34 -0.65
QCI 1 Average RSRQ -8.78 -8.62 1.82
QCI 1 Average Spectral Efficiency 1.38 1.42 2.89
QCI 1 Combined Coverage 0.38 0.42 10.52
QCI 1 Duration Seconds 210272.82 209665.94 -0.28
QCI 1 Estimated Drops 2.88 2.74 -4.86
RSRP Coverage 0.42 0.42 0
RSRQ Coverage 0.81 0.84 3.7
Spectral Efficiency Coverage 0.75 0.77 2.66
© British Telecommunications plc 2018
• RS Power (dB)
(Considers the adjustment of Reference Signal absolute Power)
• Electrical Tilt (°)
(Considers the adjustment of Electrical Tilt values for the antennas of identified cells.
All Electrical Tilt changes are between -2° and +4°)
Only 70% of proposed tilt changes were implemented, see next slide
GEOoptimize design recommendations in Cluster A
© 2018 VIAVI Confidential & Proprietary Information 19
13
119
0
50
100
150
GEOoptimize Tilt Changes Summary
Uptilt Downti lt
11
2 35
21
32
0
5
10
15
20
25
30
35
-2 -1 1 2 3 4
GEOoptimize Tilt Changes distribution
164
67
0
50
100
150
200
GEOoptimize Power Changes Summary
Power Decrease Power Increase
© British Telecommunications plc 2018
GEOoptimize design parameters (Recommended vs Implemented)
© 2018 VIAVI Confidential & Proprietary Information 20
13
119
0
50
100
150
GEOoptimize Tilt Changes Recommended
Uptilt Downti lt
8
87
0
20
40
60
80
100
GEOoptimize Tilt Changes Implemented
Uptilt Downti lt
142
60
0
50
100
150
Power Changes Implemented
Power Decrease Power Increase
164
67
0
50
100
150
200
GEOoptimize Power Changes Recommended
Power Decrease Power Increase
© British Telecommunications plc 2019
Cluster B
21
© British Telecommunications plc 2018
Cluster B executive summary
© 2018 VIAVI Confidential & Proprietary Information 22
The goal for GEOoptimize pilot was to improve :• Spectral efficiency (DL Throughput)• VoLTE Drops and Blocks
Cluster information• LTE Sites: 72• LTE Cells: 492• LTE band 1800/2600
Design Constraints:• RS power on 2600 band restricted (power increase not allowed)• RS gain for power boosting totally restricted – although this design showed significant improvement• Electrical tilt changes:
- 1800 band uptilt and down tilt restricted to -2/+4- 2600 band down tilt restricted to +4 (no up tilt allowed)- E-Tilts on shared antennas with the GSM band 1800 were not allowed
• Cell edge RXLEV/RXQUAL coverage targets for GSM1800 not availableNote: It should be noted that this was the most constrained project that Viavi has addressed with its GEOoptimize solution, duemainly to other optimization and network improvement activities that were in progress at the time.
Cluster B:Major business area in London also containing dense urban residential areas with a significant number of high rise buildings.
© British Telecommunications plc 2018
GEOoptimize LTE design prediction Cluster B
© 2018 VIAVI Confidential & Proprietary Information 23
GEOoptimize predicted that by actuating its tilt and power recommendations:
• Average Spectral Efficiency would improve by 1.4 % • Average RSRQ would increase by 1.15 %
• The estimated total number of drops would increase by 0.99 %, but QC1 (VoLTE) drops (the key KPI for us) would decrease by 3.16%
• QC1 Spectral Efficiency would improve by 1.33 %
KPI Start Value Predicted Value Delta Percentage
Average RSRP -94.15 -93.71 0.46
Average RSRQ -9.49 -9.38 1.15
Average Spectral Efficiency 1.42 1.44 1.4
Combined Coverage 0.47 0.48 2.12
Duration Seconds 3927802.75 3929868.12 0.05
Duration Seconds Variance 24050.37 25246.67 4.97
Estimated Drops 110.45 111.55 0.99
Qci 1 Average Rsrp -91.74 -91.24 0.54
Qci 1 Average Rsrq -8.95 -8.83 1.34
Qci 1 Average Spectral Efficiency 1.5 1.52 1.33
Qci 1 Combined Coverage 0.53 0.55 3.77
Qci 1 Duration Seconds 238969.79 239921.3 0.39
Qci 1 Duration Seconds Variance 208.06 209.8 0.83
Qci 1 Estimated Drops 3.16 3.06 -3.16
Qci 1 Rsrp Coverage 0.65 0.66 1.53
Qci 1 Rsrq Coverage 0.94 0.95 1.06
Qci 1 Spectral Efficiency Coverage 0.86 0.87 1.16
Rsrp Coverage 0.57 0.59 3.5
Rsrq Coverage 0.94 0.94 0
Spectral Efficiency Coverage 0.82 0.84 2.43
This table presents predictions made by GEOoptimize prior to the
design implementation
© British Telecommunications plc 2018
GEOoptimize design recommendations in Cluster B
24
Electrical Tilt (°) recommendation
• Recommended 13 cells for electrical up-Tilt
• Recommended 71 cells for electrical down-Tilt
• All Electrical Tilt changes are between -2° and +4°
RS Power (dB) recommendation
• Reference Signal (absolute power) adjustment for 164 cells
GEOoptimize predictions:
• Average Spectral Efficiency increased by 1.4%
• Average RSRQ improved by 1.15%
• VoLTE absolute Drops decreased by 3.16%
0
20
40
60
80
UpTilt DownTilt
GEOoptimize Cell Tilt Changes
0
50
100
150
-2 4 8
Part of Sector Power Changes Distribution
Power Decrease Power Increase
71
13
6
51
107
© British Telecommunications plc 2019
Results
25
© British Telecommunications plc 2018
VIAVI GEOoptimize achieved the goals agreed with EE and demonstrated value
Results
KPIs GEOoptimize design results Remarks
Cluster A
Avg. DL Throughput increased by 8.2% Step increase around 1 Mbps (from 11.84 Mbps to 12.83 Mbps)
VoLTE Drops decreased by 37.4% Decreased (from 0.202% to 0.147%)
VoLTE Blocks(Train station) decreased by 58.3%
Significant improvement (from 168 down to 70) in train station sub area. Overall QCI-1 blocking slightly improved (from 0.244% to 0.208%)
Cluster B
Avg. DL Throughput increased by 4.52% Step increase around 550 Kbps (from 12.315 Mbps to 12.865 Mbps)
VoLTE Drops decreased by 4.66% (0.213% to 0.203%) Decreased by 10.70% during working days (excluding weekend)
VoLTE Blocks Remains consistent Improved CQI 6.7%, RSRQ 1.89% and spectral efficiency 1.45% will help no degradation for VoLTE blocking
• The above optimisation goals were set up in the GEOoptimize system through easy to use web UI • With a click of a button the GEOoptimize automation process using VIAVI patented algorithms produced the design• GEOoptimize design is produced using only two hours of peak GEO located traffic from ariesoGEO platform
Pilot objectives achieved with the following constraints:
Power decrease restrictions on 2600 band
Electrical tilt changes:- 1800/800 bands e-tilt restricted to -2/+4- 2600 band allowed only down-tilt +4
RS power boost not allowed
Significant cells could not implement the design changes due to site maintenance and HW/config issues
GSM cell edge RXLEV not shared
Further improvements could potentially be achieved if constraints were to be re-considered (especially site maintenance issues)
© British Telecommunications plc 2018
• High blocking rate in two rail stations • Previously seen as a capacity issue with high user density, high volume and low throughput• GEOoptimize used tilt and power changes to reduce blocking by >50%.
Cluster A blocking hotspot example
Blocks before Blocks after
© British Telecommunications plc 2018
29© 2020 VIAVI Confidential & Proprietary Information
GEOoptimize – Business case
VIAVI propose to work with their customer to derive a business case for the GEOoptimize Energy, Voiceand Data use cases.
Energy saving:• Mobile operators can share their annual cost of energy consumption per cell/site• The PoC resulted in 2.5% cells being switched off in the small cluster without any degradation to
network performance• Extrapolating these figures nationally and considering per cell annual energy cost, should provide dollar
savings for a mobile operator.
Voice and Data:• Improved throughput, coverage/capacity and reduce blocks/drops can be translated to a dollar value.• Additionally, the improved QoE for the subscribers results in
o Customer perception and loyalty with the mobile networko Reduce churno Better network QOS, validated by the independent benchmarking reports can provide a competitive advantage
which the marketing team can benefit from.
30© 2017 Viavi Solutions Inc.www.viavisolutions.com30
© 2020 VIAVI Confidential & Proprietary Information
GEOoptimize - summary
GEOoptimize brings the power of subscriber-centric geo-located
intelligence to automated network performance
optimization
Addressing Real Business Challenges -directly maps to OpEx
savings, CapEx deferralsand customer QoE
Predictive Analytics calculate effects and show gains before changes are
made, NOT time-consuming iterative
systems - delivering more significant improvements
www.viavisolutions.com