© 2014 Qualcomm Technologies, Inc. All rights reserved.
© 2014 Qualcomm Technologies, Inc. All rights reserved.
Overview
© 2014 Qualcomm Technologies, Inc. All rights reserved.
Strong Mobile Data Demand Requires Extra Capacity
Source: Cisco VNI Mobile, 2014
0
6
12
2012 2013 2014 2015 2016 2017
Exab
ytes
per
Mo
nth
1000X
Overall Mobile Data Traffic Growth
MOBILE NETWORKS NEED TO PREPARE FOR 1000X TRAFFIC GROWTH!
0.9 EB 1.6 EB
4.7 EB
2.8 EB
11.2 EB
7.4 EB
66% CAGR 2012-2017
3
© 2014 Qualcomm Technologies, Inc. All rights reserved.
Small Cells & Extra Spectrum Are Critical For Reaching 1000x
4
MORE
SPECTRUM IN LOW AND HIGHER BANDS
MORE SMALL
CELLS EVERYWHERE!
Evolve
3G/4G/Wi-Fi
HetNets
Interference Mgmt/SON
Intelligently
Access 3G/4G/Wi-Fi
HIGHER
EFFICIENCY
MORE INDOOR
CELLS INSIDE-OUT DEPLOYMENT
© 2014 Qualcomm Technologies, Inc. All rights reserved.
Progressive Introduction Of Small Cells To Build
Dense Carrier-Grade Network
BRING CARRIER-GRADE NETWORK CLOSER
TO USER FOR NEXT LEAP OF PERFORMANCE
Macros
+ planned
small cells
+ dense
unplanned* small cells
Combined network managed by operator
* Small cells will be deployed in areas of high demand without detailed RF planning. 5
Unplanned small cell Planned small cell
© 2014 Qualcomm Technologies, Inc. All rights reserved.
Deploying Small Cells Wherever Needed
6
LOW COST DEPLOYMENT
• Minimal CapEx and OpEx
• Leverages existing premises and backhaul
EASY TO DEPLOY
• Simple plug-n-play by customer with SON
• Unplanned yet scalable and operator-managed
HIGH CAPACITY
• Significant capacity gains compared to macro-only deployment
6
Neighborhood Small Cells (NSCs)
Picocells
Enterprise small cells
Enterprise small cells
© 2014 Qualcomm Technologies, Inc. All rights reserved.
LTE OTA Network in San Diego
7
Hyper Dense Outdoor SC Network • ~15m between SCs • Semi-planned deployment • Wireless backhaul for some SCs
Dense Indoor Network • NSC with in2out coverage over
campus • Seamless mobility between cells • Unplanned deployment • Enterprise indoor network in one
of the buildings
R&D platform for UltraSON features
© 2014 Qualcomm Technologies, Inc. All rights reserved.
© 2014 Qualcomm Technologies, Inc. All rights reserved.
UltraSON Features
© 2014 Qualcomm Technologies, Inc. All rights reserved.
LTE UltraSON Features
Self Configuration: Automatic cell parameter and backhaul config.
• Automatic PCI selection
• Automatic neighbor discovery (including 3GPP
ANR)
Backhaul Aware Operation: Handle backhaul constraints
• Backhaul quality aware load balancing
Category UltraSON Features
Mobility Management: Optimize HO performance and reduce signaling load
• Frequent Handover Mitigation
• Forward handover
• Robust mobility (including 3GPP MRO)
Dynamic Resource and Tx Power Management: Optimize capacity, minimize pilot pollution and load balancing
• Tx power management
• Resource partitioning and coordination
(including 3GPP ICIC)
• Load balancing (including 3GPP MLB and
macro-SC)
© 2014 Qualcomm Technologies, Inc. All rights reserved.
SON Features Help Small Cells Deliver Carrier-Grade
Performance
• In an unplanned/semi-planned deployment, RF environment around each
small cell is different and dynamic
• Small cell needs to be able to respond when it is turned on and continue
to adapt to the changing environment
10
AT STARTUP
• Select PCI and configure neighbor list • Calibrate Tx power • Optimize idle re-selection parameters and paging
area
AFTER STARTUP
• Adapt Tx power & update neighbor list • Coordinate & partition resources with other cells • Monitor backhaul quality & prioritize preferred
users • Balance load among different cells
© 2014 Qualcomm Technologies, Inc. All rights reserved.
Automatic PCI Selection and PCI Collision Resolution
• PCI selection/reselection using Network Listen and X2 messages
• Use a combination of Network Listen, X2 messages and UE
assistance to detect/resolve any potential PCI collision caused by
hidden node
PCI 1 PCI ?
PCI 4 PCI 3
PCI 2
11
PCI 1 PCI 3
PCI 4 PCI 2
PCI 2
© 2014 Qualcomm Technologies, Inc. All rights reserved.
Automatic Neighborhood Relation (ANR)
• Target PCI to Cell ID mapping
needs to be discovered to
enable handover
– Network Listen alone cannot
detect all neighboring target
cells for handover
• Mobile reports (ANR) and X2
message exchange can be
utilized to enhance the
neighbor cell list obtained via
Network Listen
• Need to handle persistence
and dynamic nature of network
topology
Macrocell2
Macrocell1
Small Cell 1
UE
Small Cell 2
Small Cell 1 can only detect Macrocell1
UE can also detect Macrocell2 and Small Cell 2
12
© 2014 Qualcomm Technologies, Inc. All rights reserved.
Main Considerations for Mobility Management for Small Cells
Facilitate handover to small cells to maximize traffic offload
Key Issues:
Mobile UEs on small cell layer likely to cross cell boundary frequently
Excessive handovers create signaling load and potential outage and hence should be avoided
Small cell to Small cell
Small cell carrier
Overlay Macro carrier
13
© 2014 Qualcomm Technologies, Inc. All rights reserved.
Frequent Handover Mitigation (FHM) Improves User Experience
14
• Frequent handovers impact user experience and increased signaling load
• Use MRO for overall mobility parameter optimization to improve handover
performance
• Use Frequent Handover Mitigation for per UE optimization beyond MRO:
• Classify users as high speed or ping-pong users
• Adjust handover parameters per UE to prevent ping-pongs
• Handover high mobility users to macro layer
Macrocell (Frequency f1)
Small cells (Frequency f2)
Fast moving UEs crossing coverage of different small
cells
Ping-ponging UE at the edge of small cells
© 2014 Qualcomm Technologies, Inc. All rights reserved.
Reduce Handover Failures Through Mobility
Robustness Optimization (MRO)
15
• Small cells monitor handover failures and the causes
– Example: “Too Early”, “Too Late”, and “Wrong Cell” handovers
– X2 messages exchanged between source and target cells on failures
• UltraSON tracks handover failures and successes for ping-pong,
slow moving and fast moving users
• UltraSON adjusts handover parameters for each user type at each
cell to reduce handover failures
A small cell serves and optimizes mobility for users with different mobility requirements
Fast moving users
Slow moving users
Ping-pong users
© 2014 Qualcomm Technologies, Inc. All rights reserved.
Forward Handover (FHO) Improves Robustness
16
• Robust handover performance is important to ensure excellent
user experience
– E.g., for push-to-talk application
• In regular handovers, source cell would prepare target cell by
sending UE context
• If target cell did not receive UE context in time, it can fetch UE
context from the source cell to reduce handover interruption
X2 Fetch UE context
© 2014 Qualcomm Technologies, Inc. All rights reserved.
Transmit (Tx) Power Management Improves SINR
17
• 7 outdoor small cells
• Improvement in SINR with UltraSON techniques
– Median SINR improvement of 6-10 dB
UltraSON ON UltraSON OFF
© 2014 Qualcomm Technologies, Inc. All rights reserved.
Joint Tx Power and Resource Management
• Dense small cell deployment results in degradation of SINR due to inter-SC
interference
• Joint Tx power and resource management to maximize capacity and network-
wide user experience
– Resource/interference coordination to improve user SINR
– Dynamic orthogonalization of resources for cell edge users
• Use of UE measurements reports and X2 messages for Tx power and
resource management
18
All Users Benefit if the SCs use orthogonal
resources
Both Users can be served on all resources
as the interference seen is low
Network-wide fairness is increased if User B gains a
lot, in return of User A sacrificing a small fraction of
resources
A
B
Signal
Interference
© 2014 Qualcomm Technologies, Inc. All rights reserved.
• Backhaul bandwidth availability can vary due
to backhaul sharing with other traffic (e.g.,
WiFi) and other users
• Backhaul aware load balancing is done
via:
– Estimating backhaul quality by probing
bandwidth and delay estimation servers
– Offloading SC UEs to macro when SC
backhaul is the bottleneck
– Dynamic load balancing between small
cells and between small cells and macros
based on backhaul quality
– Prioritizing home/enterprise user traffic
over other user traffic
Non-cellular traffic
from backhaul owner
Internet
Neighborhood Small Cell
Residential Gateway
Competing traffic from neighbors
19
Backhaul Aware Load Balancing - Overview
© 2014 Qualcomm Technologies, Inc. All rights reserved.
Mobility Load Balancing Between Neighboring Cells
20
• Mobility Load Balancing (MLB) distributes user load between LTE
macrocell and small cells as well as between small cells to
provide better user experience and improve system capacity
• MLB uses cell load information to optimize cell boundaries to
offload users
• UltraSON uses both mobility parameters and downlink transmit
power adjustment for load balancing
A B
A B
© 2014 Qualcomm Technologies, Inc. All rights reserved.
© 2014 Qualcomm Technologies, Inc. All rights reserved.
NSC Capacity Simulation Results
© 2014 Qualcomm Technologies, Inc. All rights reserved.
Dense Urban Neighborhood Small Cells Simulation Assumptions
Red: 2 story bldg Blue: 3 story bldg
Green: 4 story bldg Cyan: 5 story bldg
Yellow: 6 story bldg
Dense-Urban Area (Simulation with Apartment Buildings)
Black star: Small cell
Magenta Circle: Mobile
Blue circle: Macrocell
met
er
Notes: a) Small cells are randomly dropped in a apartment statistically
independent of other small cells’ locations b) At most one small cell is dropped in any apartment
Parameter Value
Macrocell ISD 500m
Population Density 20000 per sq km
Number of Apartments per Macrocell
(2 subs per Apt.) 720
User Distribution
70% Indoors/
30% Outdoors;
Randomly dropped
22
© 2014 Qualcomm Technologies, Inc. All rights reserved.
Neighborhood Small Cell Capacity Simulation Dense Urban Model Configuration
• Multi-floor apartment blocks placed in a
3-cell macro area
• Each apartment block has two buildings
with a street in the middle
• 10 apartments in each floor in each
building
– Two rows of 5 apts
– Each apt is 10m x 10m with a 1m-wide
balcony
• Detailed RF propagation modeling for
indoors and outdoors
– Indoor propagation based on Keenan-
Motley multi-wall model
– Explicit modeling of internal and external
walls, windows and floor losses
• Internal wall loss: 8dB
• External wall loss: 20dB
• Floor loss: 18.3dB (indoor users only)
0 50 100 150 200 250 300 3500
50
100
150
200
250
300
350ISD=1000m
Birds-eye view
Zoomed-in layout
10m
11m
10m
Apartment Block
23
© 2014 Qualcomm Technologies, Inc. All rights reserved.
Neighborhood Small Cells Capacity Simulation Baseline (macro-only) vs. Macro+Small Cells Deployments
F1
F2
Macro Network
NSC Network
Macro + Small Cells Deployment
F1
Macro Network
Baseline Macro Deployment
Rel 8 LTE, single 10 MHz LTE carrier at 2 GHz carrier frequency
Rel 8 LTE, 10 MHz carrier for macros at 2 GHz while small cells are deployed in 3.5 GHz
24
© 2014 Qualcomm Technologies, Inc. All rights reserved.
0x
20x
40x
60x
80x
100x
120x
140x
160x
180x
0% 10% 20% 30% 40% 50%
DL
Me
dia
n T
hro
ugh
pu
t G
ain
(x)
Small Cell Penetration (%)
DL Median Throughput Gains (LTE dense urban, 10 MHz BW small cells in 3.5 GHz, gains
relative to macros only)
25 UEs/macro 200 UEs/macro
Neighborhood Small Cells Provide Significant DL Capacity
Gains for LTE
Neighborhood Small Cells Offer Scalable Capacity As Demand Increases
• 500m ISD, 720 apartments/cell, 2 subs/apartment. Users randomly dropped, 70% indoor and 30% outdoor, 2x2 MIMO • Gains shown are relative to macro-only baseline. Macros deployed in 10 MHz bandwidth at 2 GHz. Small cells deployed at 3.5 GHz with 10 MHz bandwidth. • Small cell penetration is percentage of total apartments per macrocell (720) with a small cell. For a particular operator, this number cannot exceed its own
market share. For example, an operator with 30% market share can at most have 216 small cells in a macro (assuming no small cell is deployed outside customer premise by the operator).
(360 small cells) (72 small cells)
100x capacity gains at 20% penetration
Simulation Configuration
10 MHz
@ 2 GHz
10 MHz @
3.5 GHz
Macro
Small cell
25
© 2014 Qualcomm Technologies, Inc. All rights reserved.
Exceeding 1000x Capacity Gain With Dense Neighborhood
Small Cells And More Spectrum
• 500m ISD, 720 apartments/cell, 2 subs/apartment. Users randomly dropped, 70% indoor and 30% outdoor, 2x2 MIMO • Gains relative to baseline with macros only. Macros deployed in 10 MHz bandwidth at 2 GHz. Small cells deployed at 3.5 GHz with 100 MHz bandwidth. • Small cell penetration is percentage of total apartments per macrocell (720) with a Small Cell. For a particular operator, this number cannot exceed its own
market share. For example, an operator with 30% market share can at most have 216 small cells in a macro (assuming no small cell is deployed outside customer premise by the operator).
(360 small cells)
0x
200x
400x
600x
800x
1,000x
1,200x
1,400x
1,600x
1,800x
0% 10% 20% 30% 40% 50%
DL
Me
dia
n T
hro
ugh
pu
t G
ain
(x)
Small cell Penetration (%)
DL Median Throughput Gain (LTE, dense urban, 100 MHz small cells in 3.5 GHz, relative to
macros only) 25 UEs/macro 200 UEs/macro
(72 small cells)
1000x capacity at 20% small cell penetration
Simulation Configuration
10 MHz
@ 2 GHz
100 MHz @
3.5 GHz
Macro
Small cell
26
© 2014 Qualcomm Technologies, Inc. All rights reserved.
© 2014 Qualcomm Technologies, Inc. All rights reserved.
SON Architecture: Hybrid SON
© 2014 Qualcomm Technologies, Inc. All rights reserved.
Distributed and Centralized SON Components
28
• Autonomous eNB operation with real-time decision making
– eNB utilizes rich data set obtained from the radio layers
– Reduced signaling load eases complexity and processing need at
centralized entity
• Centralized Node with long term decision making
– Receives selected metrics from RAN
– Sets policies for distributed algorithms
– Performs optimization over wide area
Sets policies, less signaling, more scalable
Distributed Decision Making
Central Node
© 2014 Qualcomm Technologies, Inc. All rights reserved.
Hybrid SON: High Level View
KPIs Alarms Filtered information
Parameter Ranges Policy Guidelines
Backh
aul
Centralized Node
RAN 2
RAN 1
© 2014 Qualcomm Technologies, Inc. All rights reserved.
Summary
• Small cells can be deployed wherever extra capacity is needed
• Hyper-dense deployment of open access neighborhood small cells utilizes additional spectrum in the most efficient way
• Small cells provide a cost-effective way for operators to reach 1000x today’s network capacity – Low-cost deployment model leveraging existing premise and backhaul
• Easy to deploy, scalable and operator-managed via SON – Qualcomm is developing UltraSON Open for robust small cell
deployment
• Hybrid SON preserves the advantages of centralized SON and distributed SON
30
© 2014 Qualcomm Technologies, Inc. All rights reserved.
Not to be used, copied, reproduced , or modified in whole or in part, nor its contents revealed in any manner to others without the express written permission of Qualcomm Technologies, Inc.
Qualcomm is a trademark of QUALCOMM Incorporated, registered in the United States and other countries. All QUALCOMM Incorporated trademarks are used with permission. Other products and brand names may be trademarks or registered trademarks of their respective owners.
© 2014 Qualcomm Technologies, Inc. All rights reserved.
research.qualcomm.com