2 - ieee wilson nokia-v2 - ieee wilson nokia-v2.pdflargenumberof live nokia networks we estimate the...

18
1 © Nokia 2017 1 © Nokia 2017 IEEE Summit Rio de Janeiro November 2018

Upload: others

Post on 10-Mar-2020

4 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: 2 - IEEE Wilson Nokia-V2 - IEEE Wilson Nokia-V2.pdfLargeNumberof Live Nokia Networks We estimate the spectral efficiency during busy hour in the busy areas from >80 live networks from

1 © Nokia 20171 © Nokia 2017

IEEE Summit Rio de JaneiroNovember 2018

Page 2: 2 - IEEE Wilson Nokia-V2 - IEEE Wilson Nokia-V2.pdfLargeNumberof Live Nokia Networks We estimate the spectral efficiency during busy hour in the busy areas from >80 live networks from

2

Spectrum assets downlink + uplink from 700 to 3600 MHz (12/2017)

0

50

100

150

200

250

Bha

rti (

MU

MB

AI)

T-M

obile

(GER

)

SpectrumMHz

Telia

(SW

E)

A1

Tele

kom

SKT

Sprin

t

Opt

us

Ora

nge

(SLO

)

AT&

T (U

SA)

Veriz

on (U

SA)

Indo

sat

T-M

obile

(USA

)

Win

d

Tele

X (

São

Paul

o)

Tele X (São Paulo):

20 MHz @ 700 MHz

5 MHz @ 900 MHz

40 MHz @ 1800 MHZ

30 MHz @ 2100 MHZ

20 MHz @ 2600 MHz

Page 3: 2 - IEEE Wilson Nokia-V2 - IEEE Wilson Nokia-V2.pdfLargeNumberof Live Nokia Networks We estimate the spectral efficiency during busy hour in the busy areas from >80 live networks from

3

LTE Spectral Efficiency in Live NetworksLarge Number of Live Nokia Networks

We estimate the spectral efficiency during busy hour in the busy areas from >80 live networks from the carried traffic per cell with a few assumptions

• 20% of BTS makes 50% of traffic• Busy hour is 7% of daily traffic• Average busy hour load is 70% of the

maximum • No voice impact considered• Average LTE bandwidth 15 MHz

1.7

1.2

0.4

2.56

1.90

0.53

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Top Top‐10% Median

bps/Hz/cell

HSPA LTE

Page 4: 2 - IEEE Wilson Nokia-V2 - IEEE Wilson Nokia-V2.pdfLargeNumberof Live Nokia Networks We estimate the spectral efficiency during busy hour in the busy areas from >80 live networks from

Tele X Mobile broadband traffic growth 2015 – 2022 per customer

2015 2016 2017

+43%

2019 20222018 2020 2021

CAGR

Page 5: 2 - IEEE Wilson Nokia-V2 - IEEE Wilson Nokia-V2.pdfLargeNumberof Live Nokia Networks We estimate the spectral efficiency during busy hour in the busy areas from >80 live networks from

5

10-20x User Data Rates with 5G up to 1 Gbps

• 200-500 Mbps average user data rates with 100 MHz 5G with mMIMO

• 400-1000 Mbps average user data rates with 200 MHz 5G with mMIMO

0100200300400500600700800900

1000

USOokla

Besturban

 Load75%

Load50%

Load25%

Load50%

Load25%

Load25%

(64TRX)

LTE 5G (100 MHz) 5G (200 MHz)

User data rate

MeasuredLTE

Simulated 5G

Page 6: 2 - IEEE Wilson Nokia-V2 - IEEE Wilson Nokia-V2.pdfLargeNumberof Live Nokia Networks We estimate the spectral efficiency during busy hour in the busy areas from >80 live networks from

6

In-Home 5G with 2.5 / 3.7 GHz – Cutting the Cord

0

1000

2000

3000

4000

5000

6000

7000

100 200 300 400 500

GB/house/month

Households per BTS

Fixed wireless capacity with 2.5/3.7 GHz

100 MHz 200 MHz• 2.5 / 3.7 GHz brings a

lot of capacity –potentially enough for fixed home access

• 1-2 TB per household per month with 300 households per BTS

• No need for directional CPE antenna

Page 7: 2 - IEEE Wilson Nokia-V2 - IEEE Wilson Nokia-V2.pdfLargeNumberof Live Nokia Networks We estimate the spectral efficiency during busy hour in the busy areas from >80 live networks from

7

5G Motivations per Band

1.8 – 2.6 GHz

Sub 1 GHz

3.3-4.9 GHz

5 GHz

24-39 GHzHotspots

Macro grid

• 5G brings mMIMO capable eco-system• 2-3x spectrum efficiency compared to LTE

• Low latency coverage for new services• Improve spectrum and energy efficiency vs LTE

• 100 MHz bandwidth for 2 Gbps• 10-20x more capacity than an LTE carrier

• Local solution without spectrum license• Improved quality compared to WiFi

• Local solution with spectrum license• Peak rate > 5 Gbps

Page 8: 2 - IEEE Wilson Nokia-V2 - IEEE Wilson Nokia-V2.pdfLargeNumberof Live Nokia Networks We estimate the spectral efficiency during busy hour in the busy areas from >80 live networks from

8

3.5 GHz Spectrum Auctions Results

• Major differences in spectrum prices at 3.5 GHz

• Spectrum price in Italy turned out to be expensive (four operators & high data usage)

0.15 0.14 0.14

0.05 0.04

0.36

0.000.050.100.150.200.250.300.350.40

Spain UK Korea Ireland Finland Italy

EUR/MHz/pop in 3.5 GHz auctions

Page 9: 2 - IEEE Wilson Nokia-V2 - IEEE Wilson Nokia-V2.pdfLargeNumberof Live Nokia Networks We estimate the spectral efficiency during busy hour in the busy areas from >80 live networks from

9

28 GHz for Capacity Boost and Offloading

US City 1

US City 2

EU City 1

81%

65%

41%

81%74%

US City 3

US City 4

Korea City 1

Hong Kong

49% 49%

Korea City 2

76%66%

Japan City 1

Outdoor Downlink Coverage (%)

4x data rate with 28 GHz (250 MHz) + 3.5 GHz (40 MHz) vs 3.5 GHz only

Dense urban outdoor coverage 40-80% with existing sites with 28 GHz

3.5 GHz only

3.5 + 28 GHz

Page 10: 2 - IEEE Wilson Nokia-V2 - IEEE Wilson Nokia-V2.pdfLargeNumberof Live Nokia Networks We estimate the spectral efficiency during busy hour in the busy areas from >80 live networks from

10

0.0

0.5

1.0

1.5

2.0

2.5

3.0

0 10 20 30 40 50

bps/Hz

PRB usage [%]

Radio efficiency benchmarking

LTE Spectral Efficiency – Global Network Wide View

• Low load typically shows>2 bps/Hz

• Very high load shows<1.7 bps/Hz

• High efficiency & highload networks mainlyfixed wireless traffic

Low load>2 bps/Hz

Data collected from network countersOne dot is one complete network

High load1.6 bps/Hz

Page 11: 2 - IEEE Wilson Nokia-V2 - IEEE Wilson Nokia-V2.pdfLargeNumberof Live Nokia Networks We estimate the spectral efficiency during busy hour in the busy areas from >80 live networks from

11

Expected Downlink Spectral Efficiency

10 MHz 2x2MIMO 2.2 1.7

20 MHz 4x4MIMO1 3.32.5

<1 GHz

1.7-2.5 GHz

100 MHz mMIMO 64x42 10.56.32.5/3.5 GHz

5GLTEBandwidth AntennasSpectrumbps/Hz/cell

20 MHz

Spectral usage

>19.6 MHz

Lean carrier

Main factors improving spectral efficiency• Massive MIMO• Device antennas• 5G enhancements150% gain assumed from 4x4 vs 2x22mMIMO gives 2.5..3.0x more capacity than 4TX

Page 12: 2 - IEEE Wilson Nokia-V2 - IEEE Wilson Nokia-V2.pdfLargeNumberof Live Nokia Networks We estimate the spectral efficiency during busy hour in the busy areas from >80 live networks from

12

Low Latency with Short Transmission Time

2 ms

0.125 ms

Minimumtransmission time

1 msHSPA

LTE

5G 1 ms

Round trip time

10-15 ms

20-30 ms

Page 13: 2 - IEEE Wilson Nokia-V2 - IEEE Wilson Nokia-V2.pdfLargeNumberof Live Nokia Networks We estimate the spectral efficiency during busy hour in the busy areas from >80 live networks from

13

Measured 5G Latency with 3.5 GHz TDD

• Measured latency 4-5 ms with 3.5 GHz TDD band with Qualcomm test device

• 3.5 GHz TDD cannot provide <3..4 mslatency due to 2.5 ms TDD frame.

• 2.5 GHz TDD cannot provide <6..10 ms latency if we need to follow TD-LTE frame structure (5 ms)

Downlink Downlink Downlink Uplink

2 ms

Page 14: 2 - IEEE Wilson Nokia-V2 - IEEE Wilson Nokia-V2.pdfLargeNumberof Live Nokia Networks We estimate the spectral efficiency during busy hour in the busy areas from >80 live networks from

14

5G Latency and Spectrum

0

2

4

6

8

10

12

LTE today Low bandFDD

3.5 GHz TDD Low bandFDD withmini‐slot

mmW

ms

Best case latency

• 5G low latency (1 ms) can beobtained with low band FDD or with mmW TDD

• 2.5/3.5 GHz TDD canprovide nice latency (<10 ms) but not ultra-low latency

5G4G

Nice latencyfor mobile broadband

Ultra-lowlatency

Page 15: 2 - IEEE Wilson Nokia-V2 - IEEE Wilson Nokia-V2.pdfLargeNumberof Live Nokia Networks We estimate the spectral efficiency during busy hour in the busy areas from >80 live networks from

15

iPhone XS Teardown

Apple Bluetooth & Wi-Fi

Application processorA12 (7 nm)

Intel basebandXMM7560 (14 nm)

Embedded SIM microcontroller

Avago high/mid bandPA

Murata 4x4 duplexer

Skyworks PAs

Intel RF transceiver

Toshiba flashmemory

Broadcomwireless charge

Intel powermanagement

Page 16: 2 - IEEE Wilson Nokia-V2 - IEEE Wilson Nokia-V2.pdfLargeNumberof Live Nokia Networks We estimate the spectral efficiency during busy hour in the busy areas from >80 live networks from

16

Cost of Connectivity in iPhones Keeps IncreasingFaster speeds, more RF bands and more antennas

• Cost of connectivityparts in iPhones havegradually increaseddue to faster speeds, more RF bands and more antennas

• iPhone XS has 1 Gbps, 25+ bands and 4x4MIMO

• 5G and mmWs bringsagain new challenges

05

101520253035404550

4 bands 3 bands 3 bands 5 bands 5 bands 10 bands 20+ bands 20+ bands 20+ bands 20+ bands 20+ bands 4x4

2G only HSPA 3.6Mbps

HSPA 7.2Mbps

HSPA 7.2Mbps

HSPA 14.4Mbps

LTE 150Mbps

LTE 150Mbps

LTE‐A 150Mbps

LTE‐A 300Mbps

LTE‐A 450Mbps

LTE‐A 600Mbps

1 Gbps

iPhone iPhone 3G iPhone3GS

iPhone 4 iPhone 4S iPhone 5 iPhone 5S iPhone 6 iPhone 6S iPhone 7 iPhone 8 iPhone XS

2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018

USD

iPhone Bill of Material (BOM) for connectivity

RF Baseband WiFi/BT/GPS

Page 17: 2 - IEEE Wilson Nokia-V2 - IEEE Wilson Nokia-V2.pdfLargeNumberof Live Nokia Networks We estimate the spectral efficiency during busy hour in the busy areas from >80 live networks from

17

5G Device Penetration based on LTE History

2012 2013 2014 2015

iPhone5 with LTE September 2012

Japan Dec 2014

LTE

5G

2016

USA March2015

Europe Dec 2016

Korea Sep 2013

UK Dec2015

2020 2021 2022 2023

iPhone with 5G

2024

LTE penetration hits 50%

5G penetration hits 50%

5G device penetration will hit 50% in most markets during 2023-2024 if we simply follow LTE history. Penetration of 30% typically 1-1.5 year earlier 2022-2023.

Page 18: 2 - IEEE Wilson Nokia-V2 - IEEE Wilson Nokia-V2.pdfLargeNumberof Live Nokia Networks We estimate the spectral efficiency during busy hour in the busy areas from >80 live networks from