PATH TO 5G Supercharging LTE & preparing mobile …the-mobile-network.com/wp-content/uploads/2016/02/Performance... · 3 ©2015 ACCEDIAN NETWORKS Company Confidential “Different

PATH TO 5G Supercharging LTE & preparing mobile networks for 5G performance

2 ©2015 ACCEDIAN NETWORKS Company Confidential

“This kind of performance leap

requires rethinking the network.”

5G VS. TODAY

KPI LTE-Rel-8* 5G (2020) Change

Traffic Volume Density

18.4 Mbps/km2 20 Gbps/km2 1000 x

Throughput 1.3 Mbps 1 Gbps 770 x

Latency 30 ms ½ ms 1-way 60 x

Availability(downtime)

95%(4320 sec/day)

99.99%(< 8 sec/day) 540 x

* Averages: NGMN White Paper (2014) & METIS / ICT-317669

3 ©2015 ACCEDIAN NETWORKS Company Confidential

“Different Apps, Different Needs”

5G ● A MULTI-SERVICE NETWORK

More

Less

Multiple services with unique needs share the same physical network

Latency Sensitive

Each will traverse its own “Network Slice” – a virtual network for each type of service.

NFV service chaining will be different for each.

Bandwidth Intense

Chatty / Bursty

High Availability

High Mobility

Loss-Sensitive

● VoLTE

● Video

● Email

●Web

● IoT – basic

● IoT – safety

4 ©2015 ACCEDIAN NETWORKS Company Confidential

“X2 latency kills throughput”

Inter-cell interference coordination (ICIC) and coordinated multi-point (CoMP) transmission require very short latencies between X2 interfaces to achieve real-time coordination between BSTs.

Source: “Centralized Scheduling for Joint-Transmission CoMP in LTE-Advanced”, S. Brueck, L. Zhao, J. Giese, M. A. Awais, proc. ITG/IEE Workshop on Smart Antennas, Bremen (WSA'10), Germany, February 2010

X2

Physical Path

Logical Path

X2

User Experience On the Line

15ms

50%

IncreasedNetwork Delay

DirectSpeed Impact

5G: LATENCY TIED TO THROUGHPUT

Delay < 1ms Required

5 ©2015 ACCEDIAN NETWORKS Company Confidential

Capacity growth can be addressed with densification

INCREASED DIVERSITY ● INCREASED DENSITY

0

1,000

2,000

3,000

4,000

5,000

6,000

7,000

8,000

CY06 CY07 CY08 CY09 CY10 CY11 CY12 CY13 CY14 CY15 CY16

Mobile Backhaul Connections by Medium

Fiber Copper Air

Source: Infonetics

x1000

Networks Soon Scaling Beyond Human Control

100 More Devices

1000 More Bandwidth

10 More Apps

1M More to Manage

By 2020, the GSMA expects:

Capacity gains will be offset by poor QoE if real-time network coordination and optimization isn’t achieved.

Standard Mobile

Cells

4G HetNet

5G: “Random”

© 2016 ACCEDIAN NETWORKS

OPTIMIZING QoE ● COMPETE ON PERFORMANCE, NOT PRICE

You Can’t Control What you Can’t See

“To gain control, providers first have to see the state of their network, from end-to-end.”

3EXPERIENCEIMPROVED CUSTOMER

COMPETITIVEDIFFERENTIATION

Visibility Affords Control

Control Optimizes Customer Experience

Providers Differentiate on Performance, Instead of Price

© 2016 ACCEDIAN NETWORKS

BUILDING THE SELF-OPTIMIZINGMOBILE NETWORK

The Three Pillars of Autonomous 5G

Networks

“Putting the mobile network at the service of each subscriber, on-demand.”

Nervous System

Ubiquitous Network State Visibility

Localized QoE Awareness

Real-Time Feedback

Responsive Control

SDN Control

Application-Aware Routing

NFV Orchestration

Decision Making

Fast Data Analytics

Deep / Machine Learning

8 ©2015 ACCEDIAN NETWORKS Company Confidential

Control Plane

Service, Location & User-Centric

Insight / Trending

VIRTUALIZED INSTRUMENTATION LAYER

Machine Learning

NFVO

SDN Control

QoS & QoE Visibility

Data Analytics

Small & Fast

Big & Deep

Subscriber ContextData Plane

Instrumentation Layer

Virtualize instrumentation for disruptive

cost and deployment efficiencyInstrumentation

Virtualized

Standards-Based PM Reflectors

NFV-Based “Modules”

Performance Assurance Controller

Director / Orchestrator

“Real-time Network State Visibility For Automated SDN & NFV Control”

9 ©2015 ACCEDIAN NETWORKS Company Confidential

5G NETWORKS ● OPTIMAL, DYNAMIC TRAFFIC PATHS

5G networks will be increasingly dynamic, crowded with competing applications, and will require exceptional QoS.

The SDN controllers required to support multi-carrier aggregation, dynamic traffic engineering and performance optimization will require a real-time feed of network performance to optimize QoE.

Performance Monitoring Required: Real-Time, Adaptive, Directional, Ubiquitous, Embedded, Open

MME / MTSO

Data Center

SDN / SON Controller

Utilization ● Available Capacity ● Packet Loss ● Latency ● Delay Variation ● QoS/E KPIs

“Intelligent Traffic Management (ITM) allows significantly higher network utilization.”

10 ©2015 ACCEDIAN NETWORKS Company Confidential

VFN LOCALIZATION FOR LATENCY OPTIMIZATION

Distributed NFV is required to locate latency-critical network functions close to the ‘users’.

Hand-offs between carriers will impact latency. RAN sharing is an option.

Performance must be assured between NFVI hosting related functions to ensure end-to-end QoE is maintained.

MME / MTSO

NFVI

Hardware-based µs measurement

precision required

< 500 µs NetworkLatency

VNF Chaining Performance

“Virtualization should improve

performance, not impact it.”

© 2016 ACCEDIAN NETWORKS

Spin, Scale, Move

NFVI

QoE-DRIVEN NFV ORCHESTRATION

Service Chains rely on responsive NFVI

Performance between VNFs is critical to QoE and the success of NFV

Real-time ‘Network State’ is the Enabler for Dynamic VNF distribution & scaling

Mobile Edge Compute (MEC) starting to address need for localization

“Scale-Up, Scale-Out and Localize ‘On Command’”

MME / MTSO

VNFs

< ms NetworkLatency

VNF Chaining Performance

VNF Catalog

NFV Orchestrator

User State

Network State+

QoS Controller

SDN Controller

VNFs

VNFs

VNFs

12 ©2015 ACCEDIAN NETWORKS Company Confidential

“You can’t control what

you can’t see.”

COMPLETE QoS & QoE VISIBILITY ● THE NETWORK STATE+

Real-time Network State & QoEvisibility needs to cover all locations and layers, with “resolution on demand” to avoid drowning in the data lake of big analytics.

fully virtualizedPerformance Assurance

VNF Controller & Director

Virtualized instrumentation can efficiently cover all corners of the RAN/EPC to provide uniform, standards-based monitoring

NFVINFV-Powered Performance Modules & Smart SFPs

Multi-Site

Multi-Flow

Multi-Layer

Multi-Vendor Standards-Based PM Reflectors

site-to-site visibility

end-to-end, segmented

visibility

EPC

© 2016 ACCEDIAN NETWORKS

FROM LTE TO 5G ● ASSURED MIGRATION

Localize performance pinch points to focus upgrades & optimization efforts

Monitor utilization trends

Tune network QoS & QoEwith real-time feedback

Instantly troubleshoot with full-layer visibility at all locations throughout the network.

“LTE-A networks can approach 5G performance with proper optimization.”

Elements

Modules

SDN Controller &NFVO, NMS/OSS

PM Actuator

Standards-Based NE PM Reflector

Performance Platform

DC

A uniform instrumentation layer for multi-vendor, multi-technology networks

Assess Readiness

Before Adding or Changing a Service

Verify Baseline at Turn-Up

Validate Roll-Out

Monitor

QoS , QoE

Sessions Tailored to Each Flow for Optimization

Derived

Signaling & Media

AssureAvailability

Actionable Fault-Management & Reporting

L2-4

L2-7

SkyLIGHTDirector

Module & VNF ControllerVCX

Active: Standards-Based+

Passive: FlowBROKER™-enabled

Test Methods:

Analyzers

14 ©2015 ACCEDIAN NETWORKS Company Confidential

THANK YOU