Mavenir Cloud RAN Solution Overview

& Roadmap
Oct-2018
Mavenir

350+ 17 of top 20
TIER 1 CARRIERS
Our Customers Support
3+ BILLION
Carriers GLOBALLY Mobile Subscribers

Industry’s
NGN Operating in
10 BILLION LEADING
150+ Monetization Requests
Annually TECHNOLOGY
COUNTRIES
With World’s First VoWiFi, VoLTE, and RCS

37% 71% 75% >50%

Share of VoLTE / Share of VoLTE / Share of RCS Share of global
VoWiFi in the U.S. VoWiFi in France, in the U.S. and messaging
and Canada Germany, UK Canada
Leading VoWiFi / VoLTE, RCS and Messaging RCS Market Share
Source: Based on GSMA, GSA, TeleGeography and Mavenir (July 2016) – Number of launched Mobile operator networks ¹ LTM as of Q3’16.

2
A Global Force

Headquarters
R&D Center

R&D Center

Other End to End Innovation for Mobile Operators

• 2,200+ employees strong with 60% R&D dedicated
• 5G Innovation Center, Israel
• AI/Security, Center of Excellence, Czech Republic
• NFV/SDN Cloud Centers: Richardson TX, India, & Shanghai
3
 Leading Innovation

Most Experienced – World First Deployments ….leading the way to 5G

NFV Multi-
VoWiFi RCS VoLTE vEPC vRCS
VoLTE ID
2011 2012 2012 2012 2013 2014 2016

Unsurpassed Large-Scale Commercial Launches Cloud Native architecture required for 5GC

Powering networks with > 75% VoLTE calls

E2E provider including innovative Cloud RAN

Most extensive Handset IOT including direct involvement

Disruptive player: technology and commercial

Mavenir Corporate Overview 4

Mavenir’s End-to-End Portfolio
New Mavenir Core Mavenir
Solutions Solutions

VAS Collect SMS / Data

Presence
Voice & Multi-ID Suite
SMS / MMS Videomail VoWiFi Server Service mStore platform:
RCS Cloud and
Creation Voice,
Messaging (Carrier and
Client Voicemail VoLTE RCS Environment Messaging, VM Interconnect Hub
Enterprise)

Access,
Voice, IMS Core
Operator Core Messaging,
Network UCC Diameter Security Suite /
IoT
EPC EIR Signaling Signaling MEG 5G NGCN
Platform
Controller Firewall

Access Session Border Security Access IoT

WebRTC / RCS API ePDG
Control Gateway Gateway

Cloud RAN
RAN

5
RFI Summary
• Demonstrate Mavenir can be a NR and eLTE partner for Telefonica
• Highlight the Mavenir technical solutions that can enable Telefonica
• Validate Mavenir’s capability around the NR environment and how
it can be applied to Telefonica
• Establish a framework in which Mavenir will support and invest in
making 5G a reality for Telefonica.
• RFI Products:
– Mavenir:
• CU & DU (vBBU), xRAN Controller S1 Aggregator, CMS (OAM), NFV
Framework
– Partner Products:
• NEC RRU
• Other vendor ecosystem.
4.5G - Evolution to 5G
Today
Mavenir – The disruptive RAN provider

Rel-15 compliant cloud RAN solutions

• Active in various forums such as 3GPP, xRAN, CBRS, ONAP, SCF

Open, standardized interfaces for a disaggregated RAN (3GPP + xRAN)

• Provide modular, end-to-end solutions for operator
• Break vendor dependence for operator

Solutions based on COTS hardware

• Flexibility in all layers to enable network slicing for 5G
• Accelerators for L1 are based on FPGAs

ONAP support -- NFV, SDN integrated into platform from day 1

• Built on a decade of virtualization experience grown at Mavenir

4G also supported via same vRAN architecture & platform

• Dual interfaces (split-2, split-7) support
• Support for the vEPC for LTE/NR with service Based Architecture

8
Mavenir RAN Development Summary

Rel-15
Open interfaces FH opt.

AVX2, Task
CPU
Mavenir
pooling, VM, opt.
containers 5G

eMBB, URLLC, New

mMTC opt. features

© 2016 Mavenir COMMUNICATIONS SYSTEMS, INC. INTERNAL USE ONLY 9

CloudRange Infrastructure (ETSI MANO)
Small cell & Macro Connectivity Architecture
Cloud RAN CloudRange Infrastructure

DU VNF
RRU OVS/SRIOV
NFVI CAS CEM VNFM

NetConf
Ve- Vi-VNF
Ethernet FH VNF
(Split 7) Midhaul
(Split 2)
CU-CP CU-UP RIC UDSF
VNF VNF VNF VNF
OVS/SRIOV
RRU
DU VNF NFVI
OVS/SRIOV
NFVI
PCIe
PHY
offload Mavenir Cloud RAN Elements

Mavenir MANO Elements

3rd Party RRUs
Operators NFV Infrastructure
DU COTS X86 Server
Kontron, Dell, HP, Lenovo ..

FPGA PCIe
H-PHY or CPRI
Adaptor

Standard X86 server:

• Minimum 2 10 Gbps interfaces (in the future up to 40 Gbps)
• Available up to 3 PCIe X16 Slots
Mavenir RRU Strategy
Expanded eco-system play with different partner RRU vendors

Investment in new 5G verticals aligned with market opportunity

•Network Slicing Mavenir Cloud RAN
•Private EPC
•CBRS
•IoT: LTE-M & NB-IoT CU
•Massive MIMO

Intel X710 NIC

DU
PCIe
H-PHY or CPRI Adaptor

Ethernet Split 7 CPRI Ethernet Split 2

Lo-PHY/RF RF
(3rd party)
RF (3rd party)
MARO over Ideal (3rd party) MICRO over ISP and
Backhaul non ideal backhaul
DU Software Implementation

DU VNF Two DU variants are available:

RRU OVS/SRIOV
• Bare metal DU
NFVI
• Can be deployed over any standard
PCIe
PHY COST X86 server with addition of
offload PCIe based PHY accelerator
• Virtual DU
Ethernet FH
(Split 7)
• Can be deployed over any standard
COST X86 server with addition of
PCIe based PHY accelerator
• Current deployments over the
RRU
WindRiver titanium Server OpenStack
DU VNF
Edge Server
PCIe OVS/SRIOV • Supports elastic scaling based on the
PHY NFVI
PCIe
offload number of RRUs
PHY
offload
Option 3x Architecture Support
Apps1 Slice Orchestration Appsn

A1 A1 A1
RIC

E2
S1-MME

Slice Specific
Slice Specific
C NFs
S/P-GW
NFs CP
RRC RRC U

MME
Split
PDCP NR-PDCP Slice Specific
Bearer Slice Specific
NFs
S/P-GW
NFs UP

Direct
Bearer Direct
SR V1 X2 F1
Bearer
B
S1-U
D
U
RLC NR-RLC
• Both core and access are built “NFV
MAC NR-MAC native” taking micro services
architecture into account
PHY NR-PHY
• Option 3x for NSA is available today for
eNB gNB lab entry

16
Option 7x Architecture Support
Apps1 Slice Orchestration Appsn

A1 A1 A1
RIC

N2
E2

CU
RRC Slice Specific
Slice Specific
NFs SMF
NFs
RRC SDAP

AMF
Split
PDCP NR-PDCP Slice Specific
Slice Specific
Bearer
NFs UPF
NFs

SRB Direct Bearer Direct Bearer

V1 Xn F1
N3
DU

RLC NR-RLC

MAC NR-MAC • Both core and access are built “NFV native”
taking micro services architecture into account
PHY NR-PHY
• Option 7x for NSA is available by Q3 for lab entry
eNB gNB

17
 CMS Position in Mavenir NFV Architecture
OSS / BSS FM PM CM

FM: PM: CM: (automated)

HTTPS - SNMP or - SFTP pull or - Netconf
SSL CLI - REST - REST - REST
AD
LDAP

webUI CMS NFVO

LIMS X1
PM: KPI, TRLs
- SFTP push/pull
- rSync CMS – CAS (CM, FM, PM)
CMS – VNF (CM, FM, PM): Netconf CMS – VNFM (LCM):
- Netconf
CDRs: E - Netconf
- SFTP push/pull M - REST (Ve-VNFM)
CAS – CMS (PM):
Application E
- rSync M
- SFTP pull Application
VNFs E
VNFs M
CU managed
CAS as VNF VNFM
VNFM – VNF (LCM)
- REST (Ve-VNFM)
VNF - CAS (CDRs/TRLs) REST E
Application M E
Application
VNFs M E
VIM
VNFs
RRUs managedM
as PNF

18
OAM

• Centralized Management
System for managing entire
Mavenir ecosystem.
• FM/CM functions by CMS.
• PM functions done by CAS
• CloudRAN OAM:
– RRUs treated as PNFs
– BBUs treated as VNFs.
• NBI interfaces presented to
OPCO OSS systems.
• Configuration Management
performed by CMS using
Netconf.
• Will be used to perform zero-
touch provisioning for
residential and Enterprise
Small cells

Insert title for presentation on master slide 19

 Open Data Model – Interfaces and Protocols
• Two additional interfaces for
OAM
– NetConf – Configuration of DU NetConf
(Alarms/Events, KPIs, Configuration)
[Mandatory] RRU
OSS CMS
– REST – Lifecycle LCM using
PNF
management of PNFs REST API
[Optional]
• NetConf Orchestration

– Alarms, Events, Configuration,

KPIs
– Data model based on TR-196
RRU
– Yang used as language BBU RRU
• REST API VNF Fronthaul Interface PNF

– Lifecycle management of DU
– Works with VNF manager
– Provisioning, termination,
monitoring, updating of PNF

Insert title for presentation on master slide 20

xRAN Based Fronthaul Solution
xRAN Fronthaul architecture
 Bandwidth reduction w.r.t. CPRI

BW
Use Case Red. Benefit
Factor
- Simple & low-cost RU design
Simple radio - More carriers per fiber over Cloud RAN
3x
2T/4T TM 1-4 fronthaul
- More efficient vRAN compute infrastructure
- Significantly reduced fronthaul BW
Massive MIMO - Reduces fiber needs from bottom to top of
64 TRX TM 7-10 tower
12x
- 16 DL Layers - Reduces fronthaul fiber needs or enables
- 16 UL Streams re-use of existing fiber in Cloud RAN
deployments
 How is the fronthaul reduction achieved?
Reduced data transport compared to CPRI
• Only 1200 out of 2048 samples are sent in frequency domain (20 MHz)
• CP is not sent on fronthaul

Intelligent quantization
• Bit-widths for UL can be adjusted based on modulation scheme and SNR

Intelligent tone selection

• Only tones that carry useful data are transmitted on fronthaul

Latency estimation and adaptation [Mavenir]

• Logic that estimates fronthaul latency and bandwidth
• HARQ latency requirements (4 ms turn-around) can be exceeded using predictive techniques (no loss seen for RTT
delays even exceeding 16 ms under low mobility conditions)
• Scheduler adjusts allocations under latency and bandwidth limitations over fronthaul

Further innovation under discussion in xRAN

Network Slicing
Technology Enablers for Slicing

• NFV
• SDN and programmable networking
• Cloud native applications
• End-to-end orchestration
• Network big data, analytics, machine learning
• Services oriented architectures

26
X2 (Xn-For 5G) Aggregator
Applications/NW orchestration

A1 (Northbound Interface)
xRANc
A1/B1 supported
Following 3GPP for
E1, F1-C, F1-U B1 (Southbound Interface)
B1

Mavenir Systems Proprietary Information

27
End-End Network eMBB
service
URLLC
service
MMTC
service

Slicing Architecture
eMBB Service to slice mapping
•Coding for large payloads
Session Mgmt
•mmWave spectrum CN CN CN
UP specific
Slice #1 Slice #2 Slice #A
URLLC
•Coding Short payload Common Network Function
•low latency – 0.1 ms TTI Orchestrator Slice #1, #2, #A
Infra
•Predictive HARQ
Service
•RACH optimization CN to RAN slice pairing
Slice
•State handling for reduced RAN/CN
signaling
RAN RAN RAN
MMTC Slice #1 Slice #2 Slice #B
•Coverage
•Large number of connections
•Unacknowledged mode RAN to Radio slice pairing
•No Handover
•Low frequency spectrum
•State handling for reduced RAN/CN
signaling RF RF 28
RF
Slice #1 Slice #2 Slice #C

Mavenir Systems Proprietary Information 28

28
Use cases
 URLLC

• Predictive HARQ - limited use of HARQ for low latency operation.

• Short TTIs (sub-ms TTI) as defined in 3GPP using mini-slots (2 symbol mini-slot as required by
3GPP NR for URLLC will be supported).

• MEC support, where the network core processing is moved to the edge.

• Dynamic resource scheduling w/ other services e.g. eMBB to reduce scheduling delay where
URLLC traffic may be transmitted on a eMBB resource.

• UL transmission scheme without grant is supported for URLLC.

• RAN slicing would be used to provide QoS related guarantees for URLLC.

30
eMBB

Mavenir is also focusing on the following solution strategies to serve eMBB

use case:
• Cloud RAN (network virtualization) using COTS, NFV ready for energy efficiency with dynamic resource scaling
• Lean carrier design: eliminate always-ON transmissions
• More dynamic beamforming with focused energy transmission to a user
• Network densification via small cells, reducing transmit power requirements while providing higher throughputs
• Improved sleep modes with ON/OFF features
• Control plane/user plane separation – allowing UP only cells (e.g. mmWave) to be turned OFF when not in use
using anchor-booster (phantom cell) concepts.

In order to meet the performance numbers for eMBB, Mavenir will support
the following technologies
• Large bandwidth support (100 MHz)
• massive MIMO support (4 simultaneous users)
• 256-QAM
• Flexible LDPC coding (for high spectral efficiency)
Roadmaps – 4G & 5G NR SOFTWARE
4G Macro Cell Solution Software (1 of 3)
2018 R 4.2 Q2 R 4.3 Q3 R 4.4 Q4
PHY: PHY: PHY:
TDD/FDD sampling rate: 7.68 PUSCH: Frequency hopping TDD/FDD sampling rate: 1.92
(5Mhz), 15.36 (10 MHz), 23.04 (15 MHz), 30.72 (20 (type 1& 2) (1.4 MHz), 3.84 ( 3 MHz)
MHz PRACH configuration High UE measurement handling:
Up to 4X4 Speed Flag RSRP, RSRQ, RSTD
Up to 256QAM in the DL PRACH & PUSCH Frequency Massive MIMO (up to 64X64)
Up to 64QAM in the UL Offset Measurements Scheduler:
Scheduler: TM8 & TM9 DL COMP
Link Adaptation: Power and Rate Beam Forming support Up to 5 Carrier Aggregation
control Scheduler: 1200 simultaneously
Proportional-fair Scheduler 700 simultaneously connected UEs in the RRC Connected state
Semi-Persistent Scheduler connected UEs in the RRC Connected state Up to 30 UEs scheduled per
200 simultaneously connected UEs Up to 20 UEs scheduled per 1ms subframe
in the RRC Connected state 1ms subframe NB-IOT
Up to 12 UEs scheduled per 1 ms UL COMP SON:
subframe RRU: Energy Savings
Up to 3 Carrier Aggregation Integration with RRU vendors Cell Outage Compensation
SON: natively supporting L-PHY split7 MLB
ANR DU with CPRI Adaptor Idle Mode Load Balancing
Automatic Physical Cell ID (PCI) SON:
configuration MRO
RACH optimization
LAB Entry Ready
4G Macro Cell Solution Software (2 of 3)
2018 R 4.2 Q2 R 4.3 Q3 R 4.4 Q4
vBBU: CU
CU
Integration with CEM (O&M) Split 2 DU-VNF CU-VNF Split
Support for horizontal scaling
Ve-VNFM Interface toward VNFM CU VNF Integration with
of CU VNF
and LCM CRDL (Session Data Store)
DU
B1 Interface towards the xRANC CU N+M HA support
PCIe based CPRI to eCPRI
Standalone CPRI to eCPRI converter with L-PHY A1 interface towards
converter integrated with DU massive MIMO
(split 7-2x) MANO/ONAP
support
L PHY – H PHY 7-2x interface DU:
DU Active/hot Standby HA
(LLS-C/U/S), M-plane (LLS-M) NetConf PCIe based CPRI to eCPRI
xRANC:
RRU: converter integrated with DU running on
Slice Management
Integration with AceAxis, Baicells COTS X86 server
over CPRI Adaptor (Cplit7) Integration with 3rd Party natively supporting
xRANC: xRAN split 7 L-PHY
Call/Bearer Admission Control xRANC:
Load Balancing and Policy
Enforcement
Radio Link Monitoring

LAB Entry Ready

4G Macro Cell Solution (4G) (3 of 3)

2018 R 4.2 Q2 R 4.3 Q3 R 1.4 Q4

Uplink MU MIMO: Uplink MU MIMO:
PHY: PHY: Optimization on Target Hardware
Non-real-time real-time:
64-Rx MMSE 64-Rx MMSE
Equalizer with IRC with max of 4-layers/UE Equalizer with IRC with max of 4-layers/UE and 4
and 4 UEs each with single layer UEs each with single layer
Rel-13 TDD version with 64T64R FDD codebook based
DL beamforming with class B feedback beamforming, legacy MIMO modes (TM1-10),
UL 64Rx MU MIMO with 4-user feedback modes of rel 8-13 including RS support
pairing, RS: CSI-RS, Class A and B
Scheduler: feedback, FDD and TDD feedback nodes as
Interference management . per rel 8-13 TM1-10 and reference signals
PFS with capacity optimization . Downlink MU MIMO:
Downlink MU MIMO: Layer mapping, modulation,
Layer mapping, modulation, subcarrier mapping, MIMO precoder for 8
subcarrier mapping, MIMO precoder for 8 layers layers and 64 Tx paths, IFFT, CP addition,
and 64 Tx paths, IFFT, CP addition, windowing, windowing, or filtering, channel coding Ph2
or filtering, channel coding Ph1
LAB Entry Ready
LTE Massive MIMO Solution Summary (Lab entry Q3 2018)

TXRUs supported for Tx and Rx: 64 (2D array Beam-width Range Supported: Horizontal: 15 RRC connected users supported: 1200
supported) deg. to 90 deg. ; Vertical: 7.5 deg. to 40 deg. Bearers Supported (DRB’s): 2400
Number of data streams (DL / UL): 16 / 8 Beam Steering support in Azimuth: -40 to +40 Simultaneous VoLTE Sessions: 300
Total Number of beams (DL / UL): 8/8 degrees Reference Signals Supported: Beamformed
Number of users per TTI : 32 Beam Steering support in Elevation: -20 to + CSI-RS, SRS,
Duplexing Mode support: FDD 20 degrees

Target max DL data rate: 1 Gbps Beam Formation and Control:

Target max UL data rate: 500 Mbps Support for multiple beamwidths according to

Min. number of carriers (20 MHz) : 2 channel type

Max Instantaneous Bandwidth: 100 MHz Beam forming for Users:

Support for Doppler: 120 Km/h Dynamic and adaptive beamforming according to
UE’s capabilities for feedback, location and
channel conditions
5G NR (1 of 2)
2018 R 5.2 Q2 R 5.3 Q3 R 5.4 Q4
PHY (Sub 6 GHz): PHY (Sub 6 GHz): PHY (Sub 6 GHz and mm-wave up to 100
OFDM Waveform for DL & UL UL control channel – ph2 MHz):
DFT-S-OFDM waveform for UL Channel coding (real-time) Subcarrier spacing and FFT size
Subcarrier spacing and FFT size (60 & 120 KHz up to 100 MHz BW)
(15 & 30 KHz up to 100 MHz BW) mini-slot based PDSCH
PRACH configuration and receiver transmission
HARQ Non-transparent DL MU-MIMO
Channel coding (non-realtime) support
UL control channel – ph1 PUSCH (multi-layer)
PUSCH non-codebook based
MIMO
PUSCH coherent MIMO

2019 R 6.1 Q1
Beam management – ph1
Beam failure recovery – ph1

PHY (Sub 6 GHz and mm-wave up to 100 MHz):

Optional PRACH configuration and
receiver
Beam management – ph2
Beam failure recovery – ph2 LAB Entry Ready
Basic DL CA operation
Basic UL CA operation
5G NR (2 of 2)
2018 R 5.2 Q2 R 5.3 Q3 R 5.4 Q4
MAC/Scheduler: MAC/Scheduler: MAC/Scheduler:
HARQ Ph1 HARQ ph2 Scheduling algorithm: PFS
Random Access procedures Ph1 Random Access proceduresPh2 HARQ: HARQ feedback
SU-MIMO: 1 UE/TTI Maintenance of Uplink Time Broadcast and Multicast:
MU-MIMO: Up to 4 UE/TTI (PAL) Alignment SIB1/RMSI, Other SI
Round Robin Scheduling Discontinuous Reception(DRX) Scalability: More than 4 UEs
Supplementary Uplink: Scheduling Handling of measurement gaps SU-MIMO: More than 1 UE/TTI
on SUL (PAL) Beam failure recovery request MU-MIMO: More than 1 UE/TTI
NSA: procedure Timing Advance
MN Terminated Split Bearer: DL Paging UL Power Control
Data Split with UM RLC Mode Procedure (PAL) NSA: Link Adaptation UL/DL
EN-DC X2 Setup: EN-DC X2 Setup MN Terminated Split Bearer: DL NSA:
Procedure (PAL) Data Split with AM RLC Mode Procedure, UL EN-DC X2 Setup: EN-DC X2
PDCP: Data Split with UM/AM RLC Mode Procedure Setup Failure, EN-DC Cell Activation
PDCP entity handling: PDCP entity MCG & SGC Bearer Procedure.
establishment SN Terminated Split Bearer SgNB Addition:
DRB: UM DRB PDCP entity handling: PDCP entity re- SgNBAdditionRequestReject
Data transfer procedures, UM Mode: Transmit establishment, PDCP entity release. PDCP:
operations: Segmentation. Receive operations: DRB: AM DRB SDU discard
Reassembly (PAL) Data transfer: Transmit operation: Status reporting: Transmit
PDCP Packet Duplication, UL Data Split operation, Receive operation.
LAB Entry Ready
Threshold. PDCP PDU routing (in case of split bearers)
5G mmWave Solution Summary
Beamforming Approach: Hybrid Beamforming Multiple access scheme: DL: OFDM with a CP Channel Coding – Channel: LDPC - physical

Antenna: 256 element array, Four 64 element UL: OFDM with a CP and DFT-s-OFDM with a layer data channels. Polar codes - physical

sub-arrays CP layer control channels, Reed-Muller codes -

Beam Steering: 4 Independently steered Duplexing: TDD smallest control payloads

beams Sub-carrier spacing - max. BW: 60 KHz - 200 CP: Normal CP. Extended CP to be supported

Carrier Aggregation support: 4 Carriers MHz, 120 KHz - 400 MHz, 240 KHz (only SS) later

Supported Physical Channels: DL: PDSCH, PDCCH, Max. FFT Size: Max 4K supported

Target Throughput values for 1 Carrier: PBCH, UL: PRACH, PUSCH, PUCCH MIMO – spatial multiplexing: DM-RS based

DL:3 Gbps, UL: 1.5 Gbps Physical Signals: DL: DMRS for closed loop, open loop and semi-open loop

PDSCH/PDCCH/PBCH, PTRS for PDSCH, CSI- transmission schemes. Both codebook and

RS, PSS/SSS, UL: DMRS for PUSCH/PUCCH, non-codebook based transmission is supported

PTRS for PUSCH, SRS At least density of CSI- in DL and UL.

RS = 3 is mandatory for mmW for beam MIMO – transmit diversity: Supported

measurement CSI: 2Tx codebook for both sub-6 MIMO – beamforming: Beam management

GHz and mmW with periodic and aperiodic beam refinement.

Modulation: DL: QPSK, 16QAM, 64QAM. Hybrid beamforming, PDSCH beam switching

UL – OFDM with CP: QPSK, 16QAM, and (optional). Periodic/Aperiodic beam report

64QAM (optional), UL – DFT-s-OFDM with CP: Beam correspondence*

π/2-BPSK, QPSK, 16QAM, and 64QAM (all Transceivers: 4

optional) MIMO layers: PDSCH: 4 layers.

PUSCH: 4 layers.
CA: Supported for up to 4 CC’s.
Thank You

40
Analyst Briefing_040418