airspan: network densification using outdoor and indoor small cells
TRANSCRIPT
2015 Small Cells Forum AirRAN: The Virtual Small Cell Moti Bitter, VP Products, Airspan Networks
2 Copyright 2015 AIRSPAN NETWORKS INC.
AirRAN “Virtual” RAN
Next Generation Radio Access Network Architecture
• Airspan Radio Access Network (AirRAN) • Holistic Virtual RAN solution: V-RAN = Distributed RAN or Centralized RAN
• same platform / same building blocks / same software… • “Small Cell centric” Software Architecture portable and vendor independent
• Uses Multi-core Network Processors (multi-vendor) • For standalone MAC and V-RAN Blade Server
• Several PHY interfaces from FAPI (aka nFAPI) to pure L1 PHY (I&Q) • Designed to be easy to port to PHY Blade Server
PHY
MAC & L3
IP interface (nFAPI)
AirRAN eNodeBs can operate in one of three modes: 1. Standalone: L1 PHY + L2/L3 MAC 2. Local Cluster: L1 PHY on Clients, L2/L3 Master 3. Centralised Cluster: Local L1 PHY
3 Copyright 2015 AIRSPAN NETWORKS INC.
AirRAN Small Cell V-RAN Hardware Architecture
V-RAN Small Cells enable Standalone, Cluster and Centralized
Single eNodeB
Local V-RAN Clusters
Network Clusters
V-RAN Centralized Baseband Blade Server
Baseband “cost optimized” for for 1 or 2 carriers, or with higher cost for 3/4 carriers Multiple baseband options per product.
EPC, HSS, PCRF and IMS
S1 Interface
Master eNodeB has higher performance baseband (runs L2/L3 for cluster), L1 Clients regular basebands
L2/L3 Master L1 Client L1 Client
S1 Interface
L2/L3 Master
L1 Client L1 Client L1 Client
Remote clusters of eNodeBs running just L1 (PHY) hosted by centralized L2/L3
Baseband server
S1 Interface
#1
#2
#3
4 Copyright 2015 AIRSPAN NETWORKS INC.
nFAPI: The Connectivity Advantage
• 2 CC CA LTE Configuration: TDD with 2x20 MHz CA, FC-2, SSF7
nFAPI requires 20 times less front-haul bandwidth And works with 8 times higher latency
when comparing with CPRI C-RAN
2x2 running 2x20 MHz CA CPRI nFAPI Fiber / Ethernet Rate 10 Gbps 1 Gbps Payload Rate ~6 Gbps ~0.3 Gbps Latency Up to 250us ≤ 2 ms Centralized-Pool HW Dedicated DSP General Num. of cells per Fiber ring (10 Gbps) 1 ~15-30
R-Units per blade 3-6 ~8-16 / 4
* Future - eCoMP
5 Copyright 2015 AIRSPAN NETWORKS INC.
• Adding street pole mounted Small Cells as a 2nd LTE layer for the Macro deployment in a dense urban area – Akihabara “Electric Town” Tokyo
• using Smart Omni Antennas
2015 SoftBank Trial – Akihabara ,Tokyo
Many thanks to SoftBank for allowing us to share some of the results!
6 Copyright 2015 AIRSPAN NETWORKS INC.
Outdoor Coverage No Load • Small Cells improves Throughput over Macro only
• Directional is slightly better than Omni
Outdoor Tests – CDF
27
41 43
Main Streets
Main Streets Avg. Throughput
Macro Only
Macro + Small Cells Omni
Macro + Small Cells Directional
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 101 106
CDF
Throughput [Mbps]
Main Inner Streets Throughput CDF
All SCs Omni With MCs
All SCs Direc>onal With MCs
Macro Only
7 Copyright 2015 AIRSPAN NETWORKS INC.
The Impact of V-RAN Small Cells: Outdoor
• Based on Outdoor Drive Tests • Note: Macro cells have 26% no coverage
• Dedicated channel utilizing 40 MHz
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110
CDF
Throughput [Mbps]
Drive Test Data (100% load)
Dedicated Channel 100% Load CoMP ON
Dedicated Channel 100% Load CoMP OFF
Macro Only 100% Load
No eICIC
8 Copyright 2015 AIRSPAN NETWORKS INC.
The Impact of V-RAN Small Cells: Outdoor
• Based on Outdoor Drive Tests • Note: Macro cells have 26% no coverage
• Co-channel utilizing 20 MHz
• CoMP improves the results by 380% for Co-Channel
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110
CDF
Throughput [Mbps]
Drive Test Data (100% load)
Co Channel 100% Load CoMP ON
Co Channel 100% Load CoMP OFF
Macro Only 100% Load
No eICIC
9 Copyright 2015 AIRSPAN NETWORKS INC.
The Impact of V-RAN Small Cells: Network Capacity
0
50
100
150
200
250
300
350
400
450
500
Macro Only Dedicated Channel Co-Channel Co-Channel w.eICIC
293
379 435
304
408 461
109
Network Aggregated Capacity (Mbps)
No CoMP (3GPP CoMP) Macro Only
• Tested across the full area • 400% higher capacity compared to Macro only
by using V-RAN Small Cells with CoMP and eICIC
x4
10 Copyright 2015 AIRSPAN NETWORKS INC.
The Impact of V-RAN Small Cells: Indoor
• Indoor shows 600% gain compared to Macro only; and (50-100)% compared to Macro + Small Cells (without CoMP)
5.1 12.1 10.2 8.8
29.7 36 35.1 34.6
75.2 72.7
62.3
48.7
1 2 3 4
Indoor Throughput Gain [Mbps]
Macros Only Macros + Small Cells CoMP ON
x6 4
3 1
2
11 Copyright 2015 AIRSPAN NETWORKS INC.
Conclusions – The value of nFAPI V-RAN Small Cells
• Rooftop Macro Cells cannot provide an adequate capacity or coverage
• In high frequency bands like 2.5 GHz and 3.5 GHz • After model tuning Coverage and Capacity can be accurately predicted
• Heterogeneous Networks are the solution to Dense Urban Networks
• An underlay of Small Cells can complete the required coverage at street level and indoor locations even in higher frequency bands:
• Small Cell can be used for coverage infill purposes in areas where macro cell layer signal is too low
• Small Cells densification increase network capacity, especially in hot spot locations.
• Small Cell gain compared to Macro only presented almost 400% outdoor gain and ~400% indoor gain
• Heterogeneous Networks Preferred Approach: • Shared Channel approach should be preferred. It appears that the wider spectrum
availability of the shared channel approach outperforms its higher interference level
12 Copyright 2015 AIRSPAN NETWORKS INC.
Conclusions – The value of nFAPI V-RAN Small Cells
• Intra Small Cell Layer Interference Mitigation – CoMP • CoMP proved to be a very effective solution for interference mitigation between Small
Cells • CoMP Aggregate Capacity Gain in Outdoor Locations observed:
• 400% compared to Macro • Outdoor Cell Edge CoMP Gain observed: 250% compared to Small Cells with no CoMP • CoMP Indoor Gain observed:
• 600% compared to Macro
• Centralized CoMP: • Centralized CoMP based on pooling L2 processing and distributed L1 processing
proved to be as efficient as pooling L1 and L2 processing using CPRI fronthaul interface
• Inter Small Cell – Macro Interference Mitigation – eICIC • Time Domain eICIC is an effective way to mitigate interference between small cell
layer and macro cell layer. • The results indicated a potential gain of eICIC in both average capacity (15%) and
cell edge performance (45%)
13 Copyright 2015 AIRSPAN NETWORKS INC.
AirRAN: Airspan’s Virtual Small Cell
2015 Winners