Spirent: Assuring the successful launch of VoWi-Fi devices and services

Assuring the Successful Launch

of VoWiFi Devices and Services

Spirent webinar

3rd June 2015

2 Spirent Communications PROPRIETARY AND CONFIDENTIAL

Speakers

Rich McNally

Director

Service Experience Programs

Spirent Communications

Rupert Lloyd

Sr. Product Manager

Service Experience Solutions


Alan Way

Sr. Manager

Methodology and Platforms



VoWiFi Overview

Alan Way


Why Voice over Wi-Fi (VoWiFi)?

Spectrum

With Mobile access exploding LTE spectrum

will be insufficient

Cost per bit: Wi-Fi spectrum is free

Over 500 MHz of spectrum in Wi-Fi

Improved Service

LTE will not penetrate large buildings very well

DAS can be used, but are expensive

Wi-Fi ideal for large office buildings, shopping

malls etc. Relatively inexpensive to deploy


Why VoWiFi?

Ubiquitous Coverage

Nearly 50M Hotspots worldwide (Jan 2015)

80% over the 2013 total

Total to hit 350M by 2018

Today 140 people for every Hotspot on

earth

By 2018 it’ll be one Hotspot per 20 people

Forecast 32M in U.S. by late 2015

# Country Hotspots

1 France 13,096,824

2 USA 9,858,246

3 UK 6,611,944

4 China 4,910,368

5 Japan 3,303,459

6 Netherlands 2,017,483

7 Belgium 2,016,641

8 Brazil 1,654,114

9 Poland 1,376,560

10 Germany 935,534

Hotspot Deployments (2014)


Next Wave of Wi-Fi

802.11ac Wave II

Nearly 7.0 Gbps of throughput

256 QAM support

MU-MIMO (4x4) supporting four spatial

streams

Support for improved Beamforming

Very attractive technology for dense UE

environments

30%

Percentage of controller-based

access points shipped in Q4 2014

that support 802.11ac (Source:

“Worldwide Quarterly WLAN

Tracker” IDC March 2015)


Single User-Multiple In Multiple Out (SU-MIMO)

DSP

RADIO

RADIO

WIRELESS ACCESS POINT


With 4x4 MU-MIMO 4 users

supported simultaneously

With 8x8 MU-MIMO 8 users

supported simultaneously

All spatial streams use the same

channel simultaneously

Acting more like a switch than a

hub

Much better spectral efficiency

Multi User-Multiple In Multiple Out (MU-MIMO)

DSP

RADIO

RADIO

RADIO

RADIO

WIRELESS ACCESS POINT


Technology and Standards Overview

Passpoint/Hotspot 2.0

Based on IEEE 802.11u

Automatic network discovery

QoS mapping

Security & Authentication EAP-SIM

AP

ANQP Query

I Support 802.11u

Network Type

Roaming Partners

Yes I can connect with this roaming partner

Partners

A

B

C

D Authentication Support

GAS


Access Network Discovery and Selection Function (ANDSF)

ANDSF

WAP

eNodeB

IP

Network


Cellular vs. Wi-Fi Data Consumption by Type

93% of U.S. data usage on tablets

via Wi-Fi vs. only 7% for cellular

Smartphones also tip to Wi-Fi by a

57-to-43% margin

MSOs projecting a 10x increase for

hotspot deployments in 2015

Feb 2014 Cisco Report forecasts

64% of cellular traffic to be

offloaded to Wi-Fi by 2018

43%

93%

7%

57%

100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%

Smartphone Tablet

Source: Adobe 2014 Mobile Benchmark Report

Cellular

Wi-Fi


Carrier Grade Wi-Fi

2010 Super Bowl in Dallas:

75GB of data per game

2015 same stadium: >5 TB

of data per game

More than 30,000 devices

During half time download

and upload can be equal


Photo Credit: James Smith /

Dallas Cowboys


Wi-Fi Calling

http://www.google.com/url?sa=i&rct=j&q=&esrc=s&frm=1&source=images&cd=&cad=rja&uact=8&ved=0CAcQjRw&url=http://logo-kid.com/sprint-logo-png.htm&ei=DApAVeXuDczuoASh5YDYCA&bvm=bv.91665533,d.cGU&psig=AFQjCNGNNxyUaoXFJJMgpSdBu7TtKcNxsw&ust=1430346630639811

http://www.google.com/url?sa=i&rct=j&q=&esrc=s&frm=1&source=images&cd=&cad=rja&uact=8&ved=0CAcQjRw&url=http://www.techweekeurope.co.uk/workspace/everything-everywhere-ee-4g-iphone-nokia-92183&ei=lsBkVcT4FsrBtQX84IDICA&bvm=bv.93990622,d.b2w&psig=AFQjCNFyNih-SkrIM76oA6WLfHO18Xdk6Q&ust=1432752598606497

http://www.google.com/url?sa=i&rct=j&q=&esrc=s&frm=1&source=images&cd=&cad=rja&uact=8&ved=0CAcQjRw&url=http://wsda.squarespace.com/endorsed-products/cellular-phone-service-verizon-wireless.html&ei=K8FkVY6sN8LfsAW_wYG4BA&bvm=bv.93990622,d.b2w&psig=AFQjCNGaKgU-Hu54jLYtg3iapz6cjzscew&ust=1432752805845139


Lessons Learned (So Far)

Rich McNally


The Experience Matters…

So Learn What Works!

The convergence dream finally a reality starting

with IMS-based VoLTE and VoWiFi

VoWiFi in the very early stages of rollout

User experience will be key to adoption

Spirent has been in the trenches, rolling out

VoWiFi with trailblazers over past yr+

Following are lessons learned for assuring a

successful launch of VoWiFi

15 PROPRIETARY AND CONFIDENTIAL


#1 – Measure What Matters

Our experience:

Functional testing is increasingly insufficient for

assuring user experience

Combined live network and lab evaluation of user

experience (UX) is a better approach

If you don’t measure everything that matters to

users, problems will emerge

Best practices:

Focus on objective assessment of what the

user experiences

Measure all key factors impacting user experience

Measure quality and consistency

Call Connection Success

Rate (Initiation & Retention)

Call Setup Time

Speech Quality

Audio Loss

Conversational Audio Delay

(Ear to Mouth Delay)

Battery Life

Key VoWiFi User

Experience Metrics


#2 – Take a Service-Centric Approach

Our experience:

A network-centric view of VoWiFi will miss key

aspects of the user experience

E.g. Handoffs can have a major impact on user

experience

Network interconnect and device interoperability

also have a major impact

Best practices:

Evaluate the service across the network

connections your customers will use

Evaluate using leading commercial devices

Measure user experience before, during and

after handoff VoWiFi – VoLTE handoffs

Wi-Fi LTE EPC/IMS

Handoff

PSTN LTE/3G

Wi-Fi


0

1

2

3

4

5

MO

S

#3 – Compare to the Alternatives (Your Customers Will!)

Our experience:

Uptake of a new service is based on how well it

works vs. existing and competitive services

Performance of competing and alternative

services is constantly changing / improving

Best practices:

Compare VoWiFi user experience to all

alternatives such as VoLTE and OTT

Compare across the key factors of user

experience that matter

Make periodic comparisons over time to stay on

top of competitor’s improvements

VoWiFi OTT

Indoor HD Speech Quality

DO

WN

LIN

K

UP

LIN

K

DO

WN

LIN

K

UP

LIN

K

DO

WN

LIN

K?

UP

LIN

K?

Actual VoWiFi and VoLTE HD speech quality

(POLQA MOS) inside an office building


#4 – Get Realistic

Our experience:

Factors such as loading and access point

performance have a major impact on experience

Packet loss and jitter due to congestion are leading

factors in degraded speech quality and

conversational speech delay

Best practices:

Measure devices and services in conditions that

accurately replicate the real world

Evaluate experience in unloaded and highly loaded

conditions

Evaluate experience in single user & multi-user

conditions

Hotspots & Enterprise Wi-Fi

Home Wi-Fi

Lots of WAPs

Lots of users

Variable load

A few WAPs

Few users

Variable load


VoWiFi Experience Evaluation Results

Rupert Lloyd


How to Measure the VoWiFi Experience?

Diagnostic App

Voice Experience Probe

Wi-Fi

LTE

Voice

Server

Device

Under Test

PSTN

LTE / 3G / Wi-Fi

Reference

Device

Voice

Experience

Probe

Variable

attenuation of

Wi-Fi signal to

control HO

Live Networks

Speech quality

Call drops / blocks

Setup time

RF Bearer

RF Levels

SSID / Cell ID

EPC/IMS


-90

-80

-70

-60

-50

Wi-

Fi R

ece

ived

Po

we

r, d

Bm

Time (s)

VoWiFi to

VoLTE

Handoff

Wi-Fi to

LTE

Handoff

VoWiFi Experience Evaluation (VoLTE-VoWiFi Handoff Example)

VoLTE Call

Starts

LTE to

Wi-Fi

Handoff

VoWiFi

Call Ends

UX Before HO UX After HO Wait

LTE HO

and cycle

ends

LTE Bearer Wi-Fi Bearer LTE Bearer

UX

During

HO


VoWiFi Experience During Handoff

Evaluation of commercial device on a

commercially available VoWiFi service

Live VoLTE and Wi-Fi network

We measured speech quality (POLQA

MOS) before and during HO

Uplink and downlink

VoLTE to VoWiFi and vice versa

0

1

2

3

4

5

1 2 3 4 5

Sp

ee

ch

Qu

ali

ty (

PO

LQ

A M

OS

)

Handoff Iterations

DownlinkBefore HO

DownlinkDuring HO

Speech During HO Speech Before HO


VoWiFi Experience During Handoff

Investigated speech quality drop

during handoff

Primary cause: RTP speech packets

lost during HO

We calculated the maximum amount

of speech lost during handoff

Uplink and downlink

VoLTE to VoWiFi and vice versa

VoLTE to VoWiFi HO had higher

average audio loss on both links

0

0.5

1

1.5

2

1 2 3 4 5 6 7 8 9 10

Up

lin

k A

ud

io L

os

t (s

)

0

0.5

1

1.5

2

1 2 3 4 5 6 7 8 9 10D

ow

nli

nk

Au

dio

Lo

st

(s)

1 2 3 4 5 6 7 8 9 10

1 2 3 4 5 6 7 8 9 10

Iterations Iterations

VoLTE to VoWiFi VoWiFi to VoLTE


VoWiFi Experience Across Networks

92%

87%

75%

80%

85%

90%

95%

WiFi-WiFi WiFi-Mobile

Call C

om

ple

tio

n S

uccess

Rate

(%

)

5.1

8.6

0

2

4

6

8

10


Call S

etu

p T

ime (

s)

More Reliable

Less Reliable

Slower

Faster

Wi-Fi to Wi-Fi was more

reliable vs. Wi-Fi to Mobile

Wi-Fi to Wi-Fi had faster call

setup vs. Wi-Fi to Mobile

Better Quality

Worse Quality

Wi-Fi to Wi-Fi had better speech

quality vs. Wi-Fi to Mobile

Call Completion Success = Successful Initiation + Successful Retention (No Drop)

Over 1,500 VoWiFi calls made in multiple locations. Two scenarios: Wi-Fi to Wi-Fi and Wi-Fi to 3G Mobile

3.3

2.5

0

1

2

3

4


Sp

eech

Qu

ality

(P

OL

QA

MO

S)


Impact of Loading on VoWiFi Experience

More Reliable

Less Reliable

Slower

Faster

Heavy load decreased

reliability by 35%

Heavy load increased call

setup time more than 2x

Better Quality

Worse Quality

HL reduced speech quality

(avg. ETM delay > 400 ms)

Unloaded Wi-Fi Access Point (VoWiFi User Only) Vs. Heavily Loaded WAP (90% of Capacity)

2.9 2.6

0

1

2

3

4

Unloaded Heavy Load

Sp

eech

Qu

ality

(P

OL

QA

MO

S)

6.9

14.4

0

5

10

15

Unloaded Heavy Load

Call S

etu

p T

ime (

s)

89%

54%

0%

20%

40%

60%

80%

100%

Unloaded Heavy Load

Call C

om

ple

tio

n S

uccess

Rate

(%

)

Ear to Mouth (ETM) Delay of > 200 ms degrades conversational speech quality


Thank you for your interest!

Watch the full webinar on the link below:

http://www.spirent.com/Assets/Video/VoWiFi

http://www.spirent.com/Assets/Video/VoWiFi