QoE Awareness for the Mobile Termination Rates Monopoly ... · QoE Awareness for the Mobile Termination Rates Monopoly Liberation Christos Tsiaras Department of Informatics IFI, Communication

© 2016 UZH, CSG@IFI

QoE Awareness for the Mobile Termination Rates Monopoly Liberation

Christos Tsiaras

Department of Informatics IFI, Communication Systems Group CSG, University of Zürich UZH

[email protected]

PhD Thesis Defense, Zurich, May 30, 2016

Introduction AQX Decompiling QoE AbaCUS MNO Selection Mechanism Summary


Mobile Call Services Ecosystem

MvNO

Alice Bob

I want to call

MTR Service rate

Total cost of the call = Service rate + MTR CPP principle → Total cost

Caller’s network

-1-


Problem

MTR since the early days of mobile communication, are considered to be a “de-facto” monopoly

Could any of you connect me with Alice?

MvNO X Alice’s MvNO

I am the only network that can connect anyone with Alice!

!  Only the MNO that Alice is connected with can reach her

!  What about multiple SIM cards? –  Not practical –  Prior agreement is needed –  Takes time –  Is Apple SIM maybe a solution?

AT&T “bans” Apple SIM!

-2-


Research Questions Answered

!  How can QoE be estimated for multiple and diverse variables?

!  How to overcome MTR monopoly? – Charging for QoE – Switch between MvNOs

!  Is an automatic and on-demand MvNO selection possible? – MvNO switching time and energy cost

-3-


AQX

!  Formalizing QoE in axioms 1.  Variables that affect QoE can be identified 2.  Those variables can be characterized

•  IVs - The more you have the better it is –  E.g., quality

•  AVs - The more you have the worst it is –  E.g., price

3.  Considering the service specifications it is possible to select the best and the worst values of each variable

4.  There is an ideal/desired/expected/agreed value (x0) for each variable 5.  The effect of each variable’s variation is identifiable

•  Influence factors (m)

6.  The importance (w) of each variable is identifiable

-4-


IV

Value [x ]

MOS[e

i(x

)]

0 x0 1

µ

e0

M

ei

x 0 − δ ′ x 0 + δ

e0 − ϵ ′

e0 + ϵ

Value [x ]

MOS[e

i(x

)]

0 x0

1µ

e0

M

m = 1m = 10−3

m = 103

QoE

IV variable

Influence factor (m)

Expected value (x0)

-5-


Influence Factors Calculation AV

Value [x]

MOS[e

a(x)]

0 x0

1µ

e0

M

x0 − δ′ x0 + δ

e0 − ϵ′

e0 + ϵ

ea

m- Area

m+ Area

m+ =

lnln 3− "ε

4ln 34

#

$

%%%

&

'

(((

ln x0 +δx0

m− =

lnln 3+ε

4ln 34

"

#

$$$

%

&

'''

ln x0 − (δx 0

QoE

IV variable

Influence factor (m)

Expected value (x0)

-6-


Generic QoE Illustration of an IV & AV

Value [x]

MOS[e

i∨a(x)]

h=

M−µ

xmin

x0

xmax

µ

e0

M

ei

ea

QoE

IV variable

Importance factor (wk)

-7-

Poor quality for free

Excellent quality at a high price

This is where

your want your

customers to be


The Design to Overcome the MTR Monopoly

!  AbaCUS –  Points to the MvNO to terminate a call –  Triggers the MvNO switching

!  Decision made by the Au2

–  The auction provide incentives to MvNOs !  Auction parameters

–  Expected QoE: MOS via AQX •  QoS-C

–  E.g., Sound quality

•  TeR-C –  E.g., MTR

!  Automatic and on-demand MNO selection mechanism –  Switches to the winning MvNO

-8-


QoS-C & TeR-C

Variable Type Classification Unit

QoS-C Sound Quality Technical IV Standard Definition (SD)

High Definition (HD)

Guaranteed Network Access Technical IV Yes No

TeR-C

MTR Economic AV Any Currency

Call Set-up Cost Economic AV -//-

Charging rate [time] Economic AV sec

Minimum Call Duration [time] Economic AV -//-

-9-


AbaCUS Algorithm

Caller MvNOs Preferences Announce TeR-Cs

No

Yes

Better option MvNO, QoS-C

Is there any better option?

Switch MvNO

Proceed without

any change

Start End

Input Condition

Output

Legend

Au2

-10-

No better option exists


MNO Selection Mechanism Implementation Steps

!  The internal Android API provides a method to select the MNO

!  Access the internal API –  Custom Android framework –  Modify the IDE access rule

!  Sign the implemented mechanism with the system signature key

!  Run the application –  System user ID –  Phone process

API

-11-


MNO Switching Mechanism Evaluation Set-up

!  Count the MNO switching time and energy consumption –  Between all possible MNO combinations –  At different times across a day –  In different locations –  In different scenarios

•  Stable in urban areas •  Moving in a train (high handover demand)

=

-12-


MNO Switching Time / Pair of MNOs

Sw −> Su Sw −> Or Su −> Sw Su −> Or Or −> Sw Or −> Su Mean

0.5

1.5

2.5

3.5

4.5

5.5

6.5

7.5

8.5

9.5

10.5

11.5

12.5

MNO switching scenarios

MN

O s

witc

hing

tim

e [s

]

MNO selection time (stable)

MNO selection time (moving)

-13-

= Or: Su: Sw:

4.3

s m

inim

um ti

me

need

ed


MNO Switching Mechanism Power Consumption Evaluation

Process Power [W]

Talk 3G 0.7050

MNO Selection moving 0.6536

MNO Selection Stable 0.5406

Talk 2G 0.3333

Consumption 2G max talk time [h] 2G power [W] 3G max talk time [h] 3G power [W]

Voice service 18.33 0.3333 8.67 0.705

Stand-by 710.00 0.0086 610.00 0.010

Ptot =%Batteryused ⋅E(J )

texperiment (s)E = 6.11Wh ⋅3600s = 21996J

PMNO(stable) =0.14 ⋅21996J

5696s= 0.5406W

-14-


Summary

!  QoE estimation while considering technical and economic variables simultaneously is possible –  With the AQX model

!  MTR monopoly obstacle can be overcome –  By charging for QoE

!  It is technically feasible to select on-demand the MvNO to terminate a call –  Because it is possible in a “reasonable” MvNO switching time

and energy cost

-15-


Conclusions

!  AQX is a generic QoE estimation model –  Multidimensional –  Non-service dependent –  Supports diverse variables simultaneously

!  Technology allows competition in the MTR market –  MTR is not a “de-facto” monopoly anymore

!  To enable competition in the MTR market an automatic and on-demand native implementation is needed –  A prototype has been implemented and evaluated for the

Android platform

-16-


Future Work

!  Incentive market analysis to reveal QoE-related variables

!  Mechanism to protect dissatisfied end-users in case of QoE is not the expected

!  Explore further auction types to increase infrastructure utilization and/or social wealth-fare

-17-


Thank you!

Q&A

QoE Awareness for the Mobile Termination Rates Monopoly ... · QoE Awareness for the Mobile Termination Rates Monopoly Liberation Christos Tsiaras Department of Informatics IFI, Communication

Documents