5G event 12 july 2016

1. “5G, Special Generation”
Professor Rahim Tafazolli
Director Institute for Communication Systems (ICS), 5GIC
Institute for Communication Systems
Home of
5G Innovation Centre

2. ICS’S STRENGTHS
Wednesday, 13 July 16 2
§ WELL KNOWN RESEARCH INSTITUTE IN
COMMUNICATIONS INTERNATIONALLY,
PARTICULARLY IN EUROPE
§ GOOD NETWORKING WITH INDUSTRY AND
ACADEMIA
§ GOOD AWARENESS OF STRATEGIC AND HOT
RESEARCH AREAS –
§ TO SOME EXTENT WE ARE THE DRIVERS

3. COMMUNICATION AND INFORMATION SYSTEMS
Wednesday, 13 July 16 3
§ MOBILE CELLULAR
§ 4G, 5G, B5G
§ WIFI
§ .11g, n, ac, ax..
§ SATELLITE
§ Broadband Fixed
§ Broadband Mobile – BB on the move
§ Broadcast – passenger vehicles
§ MACHINE TO MACHINE
§ ZigBee, LTE-M
§ VEHICLE TO VEHICLE COMMUNICATIONS
§ 802.11p, LTE-V
§ FUTURE INTERNET
§ IOT

4. University of Surrey, 5GIC
Wednesday, 13 July 16 4 r.tafazolli@surrey.ac.uk #5GIC

5. WHAT 5G IS NOT
Wednesday, 13 July 16 5
JUST MORE OF THE SAME

6. 5G is
Wednesday, 13 July 16 6
Enabler of Connected Economy and Society
WIRELESS CONNECTIVITY
BETWEEN PEOPLE AND DEVICES
FOR THE PEOPLE

7. Wednesday, 13 July 16 7
5G IN 2020 AND BEYOND
§ Enhanced Mobile Broadbandè 10/20Gbps or 1000x more capacity/area
§ Massive Machine Type Communicationsè 106 devices/km2
§ Ultra-reliable and Low Latency Communicationsè >99.999 and <1ms

8. NEW RADIO
Wednesday, 13 July 16 8

9. 1. OBE to the unavailable band between available fragments is the same as that of OFDM
2. For single carrier design 3. SFDM is 5GIC proposal
9
Figure of Merit OFDM FOFDM WOFDM FBMC GFDM UFMC SFDM3
PAPR High High High High Moderate2 High Low
OBE High Low1 Low Low Low Low Low
SE Low Low Low High High High High
ComputaLonal
Complexity
Low Moderate Moderate High High High Moderate
Short-burst traffic No No No No
Yes with the 5GIC
soluLon
Yes Yes Yes
Fragmented spectrum No No1 Yes Yes Yes Yes Yes
L1 Latency High High High High
Low with the 5GIC
soluLon
Moderate Low Low
Support for
MIMO Yes Yes Yes No
Yes with 5GIC
soluLon
Yes Yes Yes
Higher order
modulaLon
Yes Yes Yes Yes, with 5GIC
soluLon
Yes TBI Yes
NOMA Yes Yes Yes Involved Yes TBI TBI

10. SCALABLE FDM
10
SCALABLE FREQUENCY DIVISION MULTIPLEXING
10 15 20 25
10
-3
10
-2
10
-1
10
0
Es/N0
CRLB/MSE
MSE with G dep. ZC
MSE with G dep. SYM
CRLB with G
CRLB without G
5 10 15 20 25 30 35
10
-4
10
-3
10
-2
10
-1
10
0
Eb/N0 [ dB ]
BER
QPSK
16QAM
64QAM
256QAM
1024QAM
Best OBE
Lowest PAPR
CSI estimate’s
MMSE bound
approaching
Scalable
with high
order
mod
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
-70
-60
-50
-40
-30
-20
-10
0
10
Normalized frequency fs
NormalizedPSD[dB]
OFDM
TB-GFDM
HTO-GFDM
HTO-GFDM-A
OFDM
TB-SFDM
HTO-SFDM
HTO-SFDM-A
0 2 4 6 8 10 12
10
-5
10
-4
10
-3
10
-2
10
-1
10
0
PAPR in dB
CCDF
OFDM equ.10-Clus.
GFDM α=0.5 2-Clus
GFDM α=0.1 2-Clus
OFDM
TB-SFDM
HTO-SFDM

11. MU-MIMO MAP
Wednesday, 13 July 16 11
Size of MU-MIMO
NETWORKTHROUGHPUT
4x4 8x8 128 x128 1024x1024
LTE

12. MASS CONNECTIVITY
Wednesday, 13 July 16 12
Towards 1 Million per km 2
• NOVEL TECHNIQUE OF SCMA FOR IOT
• IN COOPERATION WITH HUAWEI
• 300% MORE CONNECTIVITY THAN 4G
• TRANSMISSION OF 4K ( ULTRA HIGH DEFINITION) VIDEO
ON A MOBILE NETWORK
• IN COOPERATION WITH HUAWEI AND BBC R&D
• 1ST IN THE WORLD

13. RELIABILITY, RESILIENCE & AVAILABILITY
§ EQUALLY IMPORTANT TO LOW LATENCY
§ “ALWAYS SUFFICIENT”… MULTIPLE NETWORKS AND IMPOSSIBLE WITH ONE NETWORK
§ FIXED-MOBILE CONVERGENCE
§ SPECTRUM AGILITY: ALL SPECTRUM INCLUDING LICENSED AND LICENSED-EXEMPT BANDS
§ CELL EDGE CAPACITY PROBLEM
§ NETWORK AGILITY: TRANSPORT/NETWORKING/ROUTING PROTOCOLS, SOFTWARE-BASED NETWORK ARCHITECTURE AND
SDNETWORKING
§ DISTRIBUTED CONTROL BETWEEN DEVICE AND NETWORK. NEW FLEXIBLE MAC NEEDED (MORE CONTROL TO DEVICE)
Wednesday, 13 July 16 13
>99.9999..99
RESILIENCE BY DUPLICATION AND ADDING REDUNDANCY IS NOT THE ANSWER

14. LATENCY
§ LOW LATENCIES AT ALL LAYERS, NEW PARALLEL SIGNAL PROCESSING AND COMPUTING ALGORITHMS
§ MOST DELAYS IN PREVIOUS GENERATIONS DUE TO C-PLANE SIGNALLING
§ NEW END-TO-END SIGNALLING ARCHITECTURES
§ NEW RAN
§ NEW NETWORK ARCHITECTURE (CACHING, EDGE COMPUTING..)
§ NEW APPLICATION PROTOCOLS…
Wednesday, 13 July 16 14
NOT ONLY FOR MISSION CRITICAL APPLICATION BUT MAKING NETWORKING AND ADVANCED RADIO TECHNOLOGIES WORK
EFFICIENTLY
HIGHER SPEED → does not yield┴ LOWER LATENCY
BUT
LOWER LATENCY → yields┴ HIGHER SPEED
4G INITIAL CALL SETUP TOTAL DELAY (FROM RRC IDL > CONNECTED MODE) : ~ 80 MS

15. THROUGHPUT VS. BANDWIDTH AND LATENCY (ROUND-TRIP TIME)
Wednesday, 13 July 16 15
§ BANDWIDTH CAPACITY: IDEAL THROUGHPUT A UE COULD GET:
• MAX. DATA RATE DIVIDED BY NUMBER OF ACTIVE UES IN A CELL
§ Round-trip Time (RTT): the RTT between UE and a requested content’s location can be:
• 15 to 40ms if available at eNB, BBU or P-GW
• Up to hundreds of ms if not available at network edge
§ Question: How is UE’s throughput affected by these different values?
UE eNB P-GW Content Source
10-20ms
RAN
5-20ms
Backhaul
Up to hundreds of ms
Public Internet

16. THROUGHPUT VS. BANDWIDTH AND LATENCY (ROUND-TRIP TIME)
Wednesday, 13 July 16 16
EXPERIMENT SCENARIO: SINGLE FILE DOWNLOAD
THROUGHPUT IS MEASURED USING WIRESHARK IN A CLIENT-
SERVER-BASED TESTBED (FIXED NETWORK)
§ RTT AND BANDWIDTH ARE MANIPULATED VIA NETWORK EMULATOR
§ SCENARIO: SINGLE FILE TRANSFER OVER TCP
20MBPS BANDWIDTH CAPACITY
§ LARGER DOWNLOADS (>= 5MB) ARE NOT AFFECTED MUCH BY
LATENCY.
§ SMALLER DOWNLOADS (<= 1MB) ARE NEAR-LINEARLY AFFECTED BY
LATENCY
§ AS A DOWNLOAD BECOMES SMALLER, IT IS AFFECTED MORE BY
LATENCY (I.E., THROUGHPUT DROPS QUICKER AS LATENCY
INCREASES).

17. Throughput vs. Bandwidth and Latency (Round-Trip Time)
Wednesday, 13 July 16 17
Experiment Scenario: Single File Download
100Mbps BANDWIDTH CAPACITY
§ DOWNLOAD THROUGHPUT ARE AFFECTED MORE THAN UNDER
20MBPS BANDWIDTH.
§ UNDER HIGHER BANDWIDTH, IT IS EASIER FOR LATENCY TO
BECOME THE BOTTLENECK.
§ LARGER DOWNLOADS (>= 5MB) ARE NEAR-LINEARLY AFFECTED
BY LATENCY.
§ SMALLER DOWNLOADS (<= 1MB) ARE AFFECTED MORE BY
LATENCY, ESPECIALLY AT 10-50MS (EXPONENTIALLY).

18. DATA ANALYTICS
Wednesday, 13 July 16 18
§ ENHANCE USER QOE
§ INTELLIGENT NETWORK MANAGEMENT (OSS) AND CONTENT CACHING
§ BUSINESS INTELLIGENCE
§ E-HEALTHCARE
§ AND MUCH MORE

19. Data Set & result
Wednesday, 13 July 16 19
§ RECORDS OF BBC IPLAYER: MORE THAN 100 MILLION UNIQUE REQUESTS OVER 30 DAYS
§ USING DATA OF 25 DAYS IN LEARNING PHASE
§ PREDICTING POPULARITY OF NEW CONTENTS IN NEXT 5 DAYS
§ PREDICTION ACCURACY BETTER THAN 93%

20. Operations
& Maintenance
C-RAN Baseband Units
5GIC Comm’s
Room
Ultra-Dense Cluster(2)
(AY Building)
5GIC Test-bed: Ultra-Dense Campus Test-bed
Umbrella Cell
University of Surrey Outdoor Campus
Small Cell
Ultra-Dense Cluster(1)
Macro BS
RRH
Ultra-Dense
Cluster(3)
ATI
FAP
AP
5G Soft-FDC Core
LTE Soft-Core
LTE Core (H/W)
Vodafone Reading Test Centre
VPN Tunnel
Femto Radios (x6, LTE FDD)
picoRRH’s (x6, pico-LTE TDD)
Access Points (x6, 802.11ac )
5GIC Indoor Deployment
Wednesday, 13 July 16 20

21. TESTBED EXPANSION
Wednesday, 13 July 16 21
SDN/NFV and MEC
Research,
development and
testing
Berlin
Fraunhofer
Torino, Milano, Trento, Pisa, Catania, Venezia
Telecom Italy
&
Ericsson
Guildford, Stevenage, Newbury, Basingstoke,
Brighton, Greenwich
EC Funded Project eHealth Testbed
Broadband
&
IoT
testbed

22. eHealth Testbed- Dementia
Wednesday, 13 July 16 22

23. TOPICS AND SYSTEMS OF INTERESTS
Wednesday, 13 July 16 23
5G AND BEYOND
BELOW 1GHZ, BELOW 6GHZ AND ABOVE TO TERAHERTZ
TOPICS
§ UNICAST/MULTICAST/BROADCAST
§ FIXED-MOBILE INTEGRATION
§ BROADBAND AIRINTERFACE
§ SDN/NFV/MEC/ORCHESTRATOR (SOFTWARE ARCHITECTURES AND NETWORKING)
§ ICN
§ DATA ANALYTICS
§ GEO-POSITIONING
§ LATENCY
§ RELIABILITY
§ NETWORK/DATA SECURITY
§ VERTICALS
§ E/M HEALTH
§ WIRELESS ROBOTICS
§ CONNECTED TRANSPORTS
§ ETC…
SYSTEMS
§ MOBILE CELLULAR
§ WIFI
§ SATELLITE
§ MACHINE TO MACHINE
§ VEHICLE COMMUNICATIONS
§ FUTURE INTERNET
§ IOT

24. Wednesday, 13 July 16 24
THANK YOU
CONTACT:
R.TAFAZOLLI@SURREY.AC.UK