on the benefits of 5g wireless technology for future

Sergey Shorgin

Institute of Informatics Problems of the Russian Academy of Sciences, Moscow, Russia, [email protected], http://www.ipiran.ru/english

Konstantin Samouylov, Irina Gudkova

Department of Applied Informatics and Probability Theory, Peoples’ Friendship University of Russia, Moscow, Russia, {ksam, igudkova}@sci.pfu.edu.ru, http://aipt.sci.pfu.edu.ru/en

Olga Galinina, Sergey Andreev

Tampere University of Technology, Tampere, Finland, {olga.galinina, sergey.andreev}@tut.fi, http://winter-group.net/

On the Benefits of 5G Wireless Technology for Future Mobile Cloud Computing

October 27-29, 2014

Lomonosov Moscow State University

Section "Practice and experiments in SDN & NFV"

© S. Shorgin, K. Samouylov, I. Gudkova, O. Galinina, and S. AndreevSDN & NFV 201427-29.10.2014

Major Challenges of Today

2

Increased mobile data traffic,

some say 1000x and beyond

Growth in connected devices,

up to 50 billion devices

Diverse requirements and

characteristics

Current mobile networks are likely to face capacity crunch

• a new technology that replaces 4G

• or several (integrated) technologies?

Attention shifts to what comes beyond 4G

(Fifth Generation!)


What’s in a Name?

3

Given a 10-year cycle for every existing generation,

we expect 5G systems sometime around 2020

Whereas there is currently no complete definition, 5G

may already be understood from the user perspective

Human users would like to be connected at all times

• regardless of their current location

• take advantage of services provided

by multimedia-over-wireless networks


A Glimpse of Tomorrow

4

Main challenge: user’s connectivity

experience should match service rate

requirements and be uniform

A comprehensive solution is to deploy

the higher density of smaller cells in cellular architecture

Network densification generally promises higher bit rates

and reduced energy for uplink transmission

But licensed spectrum continues to be

scarce and expensive, whereas the

traditional methods to improve its efficient

use approach their theoretical limits!


The Paradigm Shift at Work

5

We expect the majority of near-term capacity and

connectivity gains from leveraging unlicensed spectrum

Consequently, the incentive to efficiently coordinate between

the alternative radio access technologies is growing stronger

WLAN becomes an integral

part of wireless landscape

A Heterogeneous Network (HetNet)

employs hierarchical deployment

of wide-area macro cells for basic

connectivity and coverage augmented with small cells of

various footprints and by different RATs to boost capacity


Intelligent Use of Multiple Radio Access Technologies

6

Integrated cellular/WLAN deployment Own dynamic system-level simulator

• 7-cell 3GPP LTE Rel.-10 FDD

• Features diverse small cells

• Full support for IEEE 802.11-2012

• Event-driven state machine: signal

transmission, channel abstraction,

traffic and user dynamics, etc.

• Flexible statistics collection

Our focus is on dense HetNets

• Integration of cellular and WLAN

• Impact of network densification

• Advanced interference coordination

• Potential of WWAN offloading

• Energy efficient user operation


Current Picture and Perspectives

7

• Simulation-based study of multi-radio HetNets

• Dynamic stochastic geometry analysis

• Comprehensive system architecture

• Current focus on integrated deployments

• Impact of centralized vs. distributed control


Enhanced Spectral Reuse via Device-to-Device Communications

8

We study LTE/WiFi D2D offloading

• Analysis and system-level simulations

• Performance requirements and benefits

• Advanced network-assistance features

• 3GPP LTE-A & WiFi-Direct demonstration

Significant boost

in cell throughput

(up to 2x)



9

• Simulation-based study of network-assisted D2D communication

• Dynamic system analysis based on stochastic geometry

• Comprehensive architecture for D2D offloading + MWC’14 DEMO

• Current focus on emerging applications (vehicular, wearables, etc.)

• Integrating D2D as an alternative connectivity option under 3GPP


Improved Power Efficiency

10

We concentrate on energy

efficiency of a mobile device

• Optimization of Tx power per radio

• Recommendations on when each

RAT should be used

• Analysis supported by simulations

• Efficient practical control algorithms

• Framework extended to D2D & MTC



11

• Use optimization theory to solve energy efficiency problems

• Rich set of applications across HetNets, D2D, MTC, etc.

• Current focus on emerging applications

(e.g., wireless energy harvesting)

• Integrating existing energy efficient

algorithms into current networks

• Attractive trade-offs between spectral and

energy efficiencies


Efficient Support for Machine-Type Communications in LTE

12

Our goal is to improve LTE support of MTC

• Large device population w/energy constraints

• Random vs. scheduled network access

• Advanced energy/delay/success rate analysis

• Own detailed protocol-level simulator

• Efficient small data transmission mechanism

• Enhancements for idle and connected mode

Good energy

savings



13

• Comprehensive analysis of MTC overload scenario

• Efficient small data access mechanism: COBALT

• Extensive support with protocol-level simulations of 3GPP LTE

• Current focus on coexistence between MTC and H2H

• Further improvements in channel access, RRM, scalability, etc.


LTE Services and Traffic Types

14

Bitrate

Priority Example

GBR

2 Voice

4 Video

3Real time gaming

5Streaming

video

Non-GBR

1 Signalling

6,7,8,9TCP-based applications

3GPP TS 23.203: Policy and charging control architecture: Release 12. – 2014

Unicast

traffic

Multicast

traffic

GBR

Non-GBR

File size

First Come –First Served

Transparent

Processor Sharing

Elastic

traffic


Background & Teletrafficof Multiservice Loss Networks

UnicastF. Kelly, K. Ross, V. Iversen

Product form solution

Kaufman-Roberts recursion

MulticastJ. Virtamo, O. Martikainen,

K. Samouylov, Y. Gaidamaka


Recursive algorithm

ElasticT. Bonald, M. Logothetis,

G. Basharin, I. Gudkova


Recursive algorithm

15

Unicast & Multicast K. Boussetta, A.-L. Beylot, J. Virtamo,

K. Samouylov, Y. Gaidamaka


Recursive algorithm

Unicast & ElasticJ. Roberts, E. Altman, O. Boxma

No product form solution

Approximate methods

Unicast, Multicast & Elastic

(Triple Play)G. Basharin, K. Samouylov, I. Gudkova

No product form solution

Approximate methods


Radio Admission Control in LTE

Higher priority service (pre-emption capable) can get resources that were already assigned to lower priority services (pre-emption vulnerable)

RAC schemes

Partial pre-emption (service degradation)

Pre-emption capable service partially gets resources assigned to one or

more pre-emption vulnerable services (degrades services = lowers its

bit rate)

Full pre-emption (service interruption)

Pre-emption capable service fully gets resources

assigned to one or more pre-emption vulnerable services (interrupts

services)- No product form solution

- Product form solution

=

Teletraffic Models + RAC

Unicast & Multicast

Unicast & Elastic 3GPP TS 23.203: Policy and charging control

architecture: Release 12. – 2014

16


Serv

ice i

nte

rru

pti

on

Servise degradationyesno

no

yes

Unicast

traffic

Multicast

traffic

Elastic

traffic

Model №2

Service interruption Servise degradation

Model №1

Model №3

* reservation

Model №4

* threshold control

* two service

disciplines

Teletraffic Models +RAC

17

1. Borodakiy V.Y., Gudkova I.A., Samouylov K.E.,

and Markova E.V. Modelling and performance

analysis of pre-emption based radio admission

control scheme for video conferencing over LTE //

Proc. of the 2014 ITU Kaleidoscope Academic

Conference – 2014. – P. 53–59.

2. Borodakiy V.Y., Gudkova I.A., Markova E.V., and

Maslovskaya N.D. Radio admission control scheme

model with service interruption for multicast

service in LTE network // М.: PFUR. – 2014. –

P. 11–13. (rus)

3. Gudkova I.A. and Samouylov K.E. Modelling a

radio admission control scheme for video telephony

service in wireless networks // Lecture Notes in

Computer Science. – 2012. – Vol. 7469. – P. 208–

215.

4. Shorgin S.Ya., Samouylov K.E., Gudkova I.A.,

Markova E.V., and Sopin E.S. Approximating

performance measures of radio admission control

model for non real-time services with maximum bit

rates in LTE // Proc. of the 12th International

Conference of Numerical Analysis and Applied

Mathematics ICNAAM-2014. – USA, AIP

Publishing. – 2014. – Vol. 1558.


Notation Parameter

C Downlink peak bit rate, bps

, Arrival rates of requests for VC and

VoD services, 1/s 1 1,

VC and VoD service times, s

1d Bit rate for VoD service, bps

1 k Kb b b Bit rates for VC service, bps

0,1,...,n C d Number of VoD users

1, , , ,

k Km m m m State of a multicast VC session

1k

m

Session is active - VC service is

provided at least to one user on bit

rate kb

0k

m Session is not active - VC service is

not provided to users on bit rate kb

pre-emption capable pre-emption vulnerable

video conference, VC

(multicast)

Yes

(interrupt VoD)

Yes

(degraded by VoD)

video on demand, VoD

(unicast)

Yes

(degrade VC)

Yes

(interrupted by VC)

bKbK

t

Kb

d

d

d

d

d

d

d

d

inte

rruption

d

t0 t1 t2 t3 t4

d

d

d

d d bK

d

d

5, 3, 1KC b d

bK

Example of service interruption

d

d

RAC Model for Unicast and Multicast Services

18


Performance Measures

Blocking probability for video on demand

Interruption probability for video on demand

Mean bit rate for video conference

State space

1 11 ,,0 , ,0 , , , 2,,, k k knn C n C b C b n C b k Knn ee0X

1 111 1

1

, ,

C nK K

C n KKK

b bCn C

bbn C b n C

C Cp n p C

n C CC

0 0

1

1

1

1

1 10 2 1

10 2 1

, ,

, ,

k

k

k

k

C bC b K

k kn k n C b

C bC b K

kn k n C b

b p n b p n

b

p n p n

e e

e e

, ,K KB p C p C b 0 e

19


0.0002

0.0004

0.0006

0.0008

0.001

0.0012

0.0014

0.0016

0.0018

0

5 10 15 20 25 30

Inte

rrupti

on p

robab

ilit

y

for

VoD

ser

vic

e

0.1

0.3

0.5 0.7

Total offered load

3

4

5

6

7

8

0

5 10 15 20 25 30

0.1

0.3

0.5

0.7

Total offered load

Mea

n b

it r

ate

for

VC

ser

vic

eCase Study

50 MbpsC

Video Conference

1 2

3 4

1

8 Mbps, 6 Mbps,

4 Mbps, 2 Mbps,

1 hour,

/

b b

b b

Video on Demand

1

2 Mbps,

2 hour,

/

d

a

* Total offered load 1 *a Total offered load

20


RAC Model for Elastic (Non Real-Time) Services

21

t0 t1 t2 t3 t4 t5

blo

ckin

g

t

2 2 2 2 2

1 1 2 2

1 2

1 1 2 2

1 1 2 2

1, if ;

,, if

u e u e C

g u u Cu e u e C

u e u e

Notation Parameter

C Downlink peak bit rate, bps

1 2, Arrival rates of elastic flows, 1/s

1 2, Elastic flow mean sizes, bits

1 2e e Max bit rates, bps

1 2,u u Number of elastic flows in the system

U Maximum number of ongoing flows (threshold)

1 2( , )g u u Degradation factor

Example of service degradation


QoS Parameters

Blocking probability

Mean transfer time for first elastic flow

Mean transfer time for second elastic flow

State space

1 2

1 1 2

, 1 0

1

1 1

, ,

1

U u

uu u u i

u p u u iP i u i Q u

TB G Q U

X

1 2 1 2 1 2, : 0, 0,u u u u u u U X

1 2 1 2

1 2, :

,u u u u U

Q UB p u u

G

X

1 2

2 1 2

, 1 0

2

2 2

, ,

1

U u

uu u u i

u p u u u i P i u i Q u

TB G Q U

X

where conditional probability that the system is in state

given that the number of ongoing elastic flows is 𝑢𝑃𝑢 𝑢1, 𝑢2 𝑢1, 𝑢2

22


Licensed Shared Access (LSA) (1/2)

Motivation behind the LSA Technology

Modern methods of constructing and organizing cellular

networks might become inefficient under the growth of traffic

volume (Cisco forecasts: in 2014 – 2.6 EB per month, in 2018

– 15.9 EB per month)

Allocating new spectrum for constant use by cellular networks

is difficult due to the approved frequency allocation plan

However, the State can allocate unused spectrum for a short

term (month, week, etc.)

23

Cisco VNI: Global Mobile Data Traffic Forecast Update, 2013–2018



Role of the LSA Technology

LSA – the promising technology of temporary resource

allocation that allows:

an operator to lease additional frequency bands in order to

satisfy the increasing data rate requirements

the State to lease out unused frequency bands and to gain

income

24



Example Usage of the LSA Technology

25

The owner of LSA band – is airport

This band is seldom used by the

airport

Cellular operator can lease it in order

to enhance its network performance

However, the operator will need to

”return” the leased band to airport if

one requests it.

Main band is available

LSA band is available

User


LTE Cell Resource Allocation ModelBased on LSA Technology

26

2

1

C1

μ

μ

C2

μ

μ

λ1 2...

β

α

r

Notation Parameter

1C Main band downlink peak bit rate, Mbps

2C LSA band downlink peak bit rate, Mbps

Arrival rate of user requests for service , 1/s

1 Service time, s

r Buffer size (r<C2)

Rate of LSA band revocation, 1/s

Rate of LSA band restoration, 1/s

0,1, ,n r Number of users in the system

0,1s LSA band state (available/unavailable)

Taking into account all the special features of LSA technology,

we propose LTE cell model with one base station and two

bands – main band and LSA band

LSA band is available at certain moments


LSA Band Availability Model (1/7)

27

... С

...

изъятие

возвращение в

рабочее состояние

Полоса

доступна Полоса

недоступна

r

LSA band

is available

LSA band

is unavailable

Notation Parameter

C Downlink peak bit rate, Mbps

Arrival rate of user requests for service , 1/s

1 Service time, s

r Buffer size (r<C)

Rate of LSA band revocation, 1/s

Rate of LSA band restoration, 1/s

0,1, ,n r Number of users waiting to receive service

0,1, ,m C Number of users receiving service

LSA band availability model describes LSA band revocation and

restoration mechanisms

LSA band can be revoked only when the system is empty



28

Request

arrives

Service terminates

ON ON ONOFF OFFOFFON

Waiting time

Service time

Service delay

Service

starts

Service unavailability

+

Example of User Request Service



29

System State Space

System state space satisfies the relation

It could be simply obtained that the process representing the

system states is a non-reversible Markov process and the

solution to the equilibrium equations is not of

product form

We obtained a recursive algorithm for calculating the stationary

probability distribution

( , ) : (0, ), 0,..., ; ( , ), 1,..., ; ( , 0), 1,..., (1)n m m m C n C n r C n n r X

( , ), ( , )p n m n m X



30

Performance Measures

, , 0 (2)B p r C C p r

1

, 0 (3)r

n

F p n

1 1

, , 0 (4)r C r

n n

N np n C np n

Service request blocking probability by a user

Probability that the LSA band is unavailable

Mean number of users, which suffer from unavailable service

(the mean number of user requests in the buffer):



31

Service Unavailability

,N N N

N

N

,n m

Mean number of users, which suffer from unavailable service

– mean number of users, who wait until their service starts,

– mean number of users, who suffer from service interruption due to the LSA

band unavailability

For each state it is impossible to determine how many of

n users with unavailable service are the users waiting for the

service start and for how many the service has been interrupted

We employ a probabilistic method to establish the

characteristics and N N



32

Numerical Example

0

0.0002

0.0004

0.0006

0.0008

0.001

10 20 30 40 50 60 70 80

0.1

0.2

0.3

0.4

0.5

0

0.5

1

1.5

2

2.5

3

500 1000 1500 2000 2500 3000

100 MbpsC 250r

1

1

1

1 month

30 minutes,

10 minutes,

/

1

1

10 days, 10 days, 1 day,

10 minutes,

30,

/

Total offered load Restoration time

Mea

n n

um

ber

of

use

rs

awai

tin

g s

erv

ice,

N1



33

Simulation Results100 MbpsC 250r

1

1

1

1 month

30 minutes,

10 minutes.

Total offered load

Pro

bab

ilit

y t

hat

th

e sy

stem

is

emp

ty

on the benefits of 5g wireless technology for future

Documents