Mobility Protocols and Handover Optimizationsites.ieee.org/denver-com/files/2018/05/Boulder-DL-Talk.pdf2018-05-06 · 20 Types of 5G Applications Enhanced Mobile Broadband - Mobile

Mobility Protocols and Handover Optimization

Speaker: Ashutosh Dutta, Ph.D.

Founding Co-Chair – IEEE 5G Initiative

Director Industry Outreach

New Jersey

Email: [email protected]

Phone: 908-642-8593

April 16 2018, Denver Section, ComSoc Chapter

mailto:[email protected]

Copyright 2015 © IEEE. All rights reserved. 2

Talk Outline

Motivation for Optimization

Mobility protocols for multimedia

Systems analysis of mobility events

Modeling Mobility Handoff

Mobility Optimization Techniques

Sample Use Cases

Best Current Practices for Handoff Optimization


What are Characteristics of Next Generation Networks?

Heterogeneous networks, many access networks

–Access-independent converged IP network

Order-of-magnitude increases in bandwidth

–MIMO, smart antennas

–Increase in video and other high bandwidth traffic

New terminals

New services and service enabling platforms

Large range of cell sizes, coverage areas

–PAN, LAN, WAN

–Pico-cellular, micro-cellular, cellular

Densification of Cells

Changes in traffic and traffic patterns

–Rise in video on demand? Requires good high-bandwidth multicast


Evolution of wireless access technologies

2020

20 Gbp/s DL SpeedF-OFDM, SCMA



Mobile Wireless Internet: A Scenario

802.11a/b/g

Bluetooth

IPv6

Network

UMTS/CDMA

Network

InternetDomain1

Domain2

UMTS/

CDMA

PSTN gateway

Hotspot

CHRoaming

User Ad Hoc

Network

PAN

LAN

WAN

WAN

LAN

PSTN

802.11 a/b/g


Handover Taxonomy

Inter-subnet

Intra-subnet

Intra-tech &

Inter-domain

Intra-tech & Intra-domain

Inter-tech &

Inter-domain

Inter-tech &

Intra-domain

Intra-tech &

Intra-domain

802.11 (provider X) to CDMA (provider X)

802.11 (provider X) to CDMA (provider Y)

802.11b (provider X) to 802.11n (provider X)

802.11b (provider X) to 802.11n (provider Y)

Inter-tech & Intra-domain

802.11 (provider X) to CDMA (provider X)Some scenario could be homogeneous as well, e.g., intra-tech &

intra-domain


Use Case: Using Multiple Radios

Ne

tw

or

k

Ty

pe

S

SI

D/

C

ell

ID

B

S

SI

D

Op

er

at

or

Se

cu

rit

y

N

W

C

ha

nn

el

Q

o

S

Ph

ysi

cal

La

yer

Dat

a

Rat

e

GS

M

13

98

9

N/

A

AT

&T

NA NA 1

9

0

0

N

/

A

N/A 9.6

kbps

80

2.1

6d

N

A

N

A

T-

Mo

bile

PK

M

EAP-

PEA

P

1

1

Y

e

s

OF

DM

40

Mbp

s

Wakeup WLANDownload over WLANShutdown GPS

Café

Airport

Zone 1 Zone 2 Zone 3

Zone 4 Zone 5 Zone 6

Zone 7 Zone 9

Wi-Fi

Wi-MAX

WLAN Link Going Down.

Switch to WiMAXDownload over WiMAXShutdown WLANWakeup GPS Zone 8

Wi-Fi

Connect to WLAN

Battery level lowShutdown WiMAXDownload over GSM/GPRS

Wakeup WLAN

Wi-MAX

Shutdown GPSStart Download over WLAN

Network

Type

SSID/

Cell ID

BSSID Operator Security NW Channel QoS Physical

Layer

Data Rate

GSM 13989 N/A AT&T NA NA 1900 N/A N/A 9.6 kbps

Network

Type

SSID/

Cell ID

BSSID Operator Security NW Channel QoS Physical

Layer

Data Rate

GSM 13989 N/A AT&T NA NA 1900 N/A N/A 9.6 kbps

802.11b Café 00:00:… Café .11i EAP-

PEAP

6 .11e OFDM 11 Mbps

Network

Type

SSID/

Cell ID

BSSID Operator Security EAP

Type

Channel QoS Physical

Layer

Data Rate

GSM 13989 N/A AT&T NA NA 1900 N/A N/A 9.6 Kbps

802.11b Airport 00:00:… Airport .11i EAP-

PEAP

6 .11e OFDM 11 Mbps

Radio State

GSM

WLAN

WiMAX

GPS

Radio State

GSM

WLAN

WiMAX

GPS

Radio State

GSM

WLAN

WiMAX

GPS

Radio State

GSM

WLAN

WiMAX

GPS

Radio State

GSM

WLAN

WiMAX

GPS

Radio State

GSM

WLAN

WiMAX

GPS

Radio State

GSM

WLAN

WiMAX

GPS

802.21 and MP Enabled Seamless Mobility Deployment Scenario

Courtesy: Vivek Gupta, IEEE 802.21 chair


Non-optimized handoff results

Handoff between heterogeneous access

(802.11 – CDMA)

Handoff between homogeneous access

(802.11 – 802.11)

c. SIP-based non-optimized

handoff between 802.11 networks

802.11 802.11Handoff

Delay 4 s

Handoff Delay

~ 18 s

802.11 CDMA

Handoff Delay

16 s

802.11 CDMA

a. MIP-based Non-optimized handoff

b. SIP-based Non-optimized handoff


Multiple Interface Case (802.11b – CDMA1XRTT) – MIP as mobility protocol

802.11 802.11CDMAHandoff19 s

Effect of handoff delay during non-optimized mobility management (experimental results)

Single Interface Case (802.11b – 802.11b) – SIP as mobility

802.11 802.11Handoff

4 s

Handoff

17 s802.11 CDMA 802.11

Multiple Interface Case (802.11b – CDMA1XRTT) – SIP as mobility protocol


Motivation for OptimizationHandoff contributes to– Change in network connection path between communicating nodes– Discrete Sate Event change at different layers– Rebinding of common set of properties (e.g., association, endpoint address,

locator)– Associated delay and packet loss due to these discrete events and rebinding

Limit jitter, delay and packet loss for real-time applications during different types of handoff

– 150 ms end-to-end delay and 3% packet loss for interactive traffic such as VoIP– ITU-T G.114

Essential to reduce handoff delay across layers during re-association and mitigate the effect of handoff delay (i.e., packet loss)– Currently it takes between 4s – 17 s– Packet loss depends upon the CODEC, packet generation rate (G711, G729)

The challenge is even greater when moving between– Heterogeneous domains – Heterogeneous access technologies (e.g., CDMA, 802.11)– Simultaneous mobility


Cellular mobility typically involves handoff across homogeneous access technology – Optimization techniques are carefully engineered to

improve the handoff performanceIP-based mobility involves movement across access technologies, administrative domains, at multiple layers and involve interaction between multiple protocols– Mechanisms and design principles for optimized

handover are poorly understood– Currently there are ad hoc solutions for IP mobility

optimization, not engineering practice – No formal methodology to systematically discover or

evaluate mobility optimizations – No methodology for systematic evaluation or

prediction of "run-time" cost/benefit tradeoffs

12

Optimization for IP-based Mobility


Evolution of Mobility Protocols


1G Cellular Architecture


GSM-based Mobility Architecture (2G)


WCDMA Cellular Architecture (3G)


CDMA 2000 Architecture (3G)


LTE 4G Deployment Scenario

HSS

AAA

PGW (PCEF)MME

ePDG

PCRF

eNodeB eNodeB

SGW

SGSN S10

S11 S11S4

S3

S8Gx

RxS6aSGi

S2b

SWx

X2

S1-U

App Servers

S6b

DNS

ENUM

Non-LTE Access (WiFi)

• Monitoring traffic

at control and user

planes

• Monitoring tunnels and

pair performance

• Correlating traffic to mobile device

• Handover and roaming

• Registration/Admission Control on AAA server

• IP multimedia services

Data Center/IMS

Mobile Core/EPC

Access & Backhaul

UE

Monitoring PointSGW HSGW

MME

I-CSCF

Internet

Gm

S1-MME

3G Access (UTRAN)

RNC

P-CSCF

Cx

S-CSCF

Cx

Mw

LTE Access

Mw

S2a

Mw

Trusted non-LTE Access (EV-DO)

S103

S6d

SGi

Iu-psS12

GGSN

Gn

Gm

http://www.google.com/aclk?sa=L&ai=CtFoU1htZS9DHA4z28AaCoLjFCobdoXzmnsS5DO7ko64wCAIQAaoEBU_QDIiG&sig=AGiWqtwy7EpB5evk2Bkxzr2y1t1NLdwZRQ&q=http://clickserve.cc-dt.com/link/prdad?lid=41000000015969363&pid=49817812&pubid=21000000000232270&mid=ChMI0Ir8h4e3nwIVDDvcCh0CEK6oEAIYASAA

http://www.google.com/aclk?sa=L&ai=CtFoU1htZS9DHA4z28AaCoLjFCobdoXzmnsS5DO7ko64wCAIQAaoEBU_QDIiG&sig=AGiWqtwy7EpB5evk2Bkxzr2y1t1NLdwZRQ&q=http://clickserve.cc-dt.com/link/prdad?lid=41000000015969363&pid=49817812&pubid=21000000000232270&mid=ChMI0Ir8h4e3nwIVDDvcCh0CEK6oEAIYASAA

• D2D Communications

• Efficient Small Data Transmission

• Wireless Backhaul / Access Integration

• Flexible Networks

• Flexible Mobility

• Context Aware Networking

• Information Centric Networking

• Moving Networks19

Key Characteristics of 5G

• Massive MIMO

• RAN Transmission –Centimeter and Millimeter Waves

• New Waveforms

• Shared Spectrum Access

• Advanced Inter-Node Coordination

• Simultaneous Transmission Reception

• Multi-RAT Integration & Management

20

Types of 5G Applications

Enhanced Mobile Broadband

- Mobile Broadband, UHD / Hologram, High-mobility, Virtual PresenceCritical Communications

- Interactive Game / Sports, Industrial Control, Drone / Robot / Vehicle, EmergencyMassive Machine Type Communications

- Subway / Stadium Service, eHealth, Wearables, Inventory ControlNetwork Operation

- Network Slicing, Routing, Migration and Interworking, Energy SavingEnhancement of Vehicle-to-Everything

- Autonomous Driving, safety and non-safety features

Massive Sensing

1b/s over 10 years

off an AAA battery

Speed: >10 Gb/s Tb/s

Massive Content

Massive Control

Response: 1 msCourtesy: Gerhard Fettweis

21

What “5G and Advanced Communication Systems” is About


5G Eco System

Source 5GPP


Mobility TaxonomyIP Mobility

PersonalTerminal Service

Application

Layer

Network

Layer

Session

• Systems

Optimization

MIPv4 CIPHAWAIIIDMP MIP-LR MIPV6ProxyMIPv6

SIPMM

MIP-LR(M)

Proxy

Transport

Layer

MSOCKS,

Migrate

mSCTP

Shim

Layer

HIP

Issues

• Host controlled

vs.

Mobile Controlled

• Mobility pattern




ForeignSubnet

ForeignSubnet

Hierarchical Mobile IP

IP-based Network

CH

HomeSubnet

HA

<CH.IP, MH.IP>

<MH.IP, CH.IP>

MH

RFA

CH to MH

CH sends packet to MH home address as usual

HA in home subnet intercepts packet, tunnels it to GFA

GFA un-encapsulates packet, tunnels it to RFA

RFA un-encapsulates packet, sends to MH

home

network

GFA coverage area

GFA

RFA


SIP-based Mobility – Application Layer


Network-based Mobility


Backbone

Administrative

Domain A

L2 PoA

Corresponding

Host

128.59.10.7

IPch

207.3.232.10

207.3.240.10

128.59.11.8

N2

N1N1

N2

N1- Network 1 (802.11)

N2- Network 2 ( CDMA/GPRS)

Configuration

Agent

L3 PoA

207.3.232.10

Mobile

Host

Authentication

Agent

Authorization

AgentRegistration

Agent

Registration

Agent

Administrative

Domain B

Configuration

Agent

Authorization

Agent

Signaling

Proxy

Authentication

Agent

Signaling

Proxy

L3 PoA

L2 PoA

L2 PoA

L2 PoA

L2 PoA

L2 PoA

L3 PoA

Mobility Illustration in IP-based 4G network

128.59.9.6

900 ms

900 ms

802.11 802.11

802.11802.11Handoff

Delay 4 s

4 seconds

Handoff Delay

~ 18 s

802.11 CDMA

18 seconds


Mobility/

Function

Access

Type

Network

Discovery

Resource

Discovery

Triggering

Technique

Detection

Technique

Configuration Key exchange/

Authentication

Encryption Binding

Update

Media

Rerouting

GSM TDMA BCCH FCCH Channel

Strength

SCH TMSI SRES/A3 DES MSC

Contld.

Anchor

WCDMA CDMA PILOT SYNC

Channel

Channel

Strength

Frequenc

y

TMSI SRES/A3 AES Network

Control

Anchor

IS-95 CDMA PILOT SYNC

channel

Channel

Strength

RTC TMSI Diffie-

Hellman

AKA

Kasumi MSC

Contld.

Anchor

MSC

CDMA

1X-

EVDO

EVDO PILOT

Channel

SYNC

Channel

Channel

Strength

RTC TMSI Diffie-

Hellman/

CAVE

AES MSC PDSN/MSC

802.11 CSMA/

CA

Beacon

11R

11R

802.21

SNR at

Mobile

Scanning.

Channel

Number,

SSID

SSID,

Channel

number

Layer 2

authenticate

802.1X

EAP

WEP/WPA

802.11i

Associate IAPP

Cell IP Any Gateway

beacon

Mobile

msmt.

AP

beacon

ID

GW

Beacon

MAC

Address

AP address

IPSec IPSec Route

Update

Intermediate

y

Router

MIPv4 Any ICMP

Router

adv.

FA adv.

ICMP

Router

Adv.

FA adv.

L2

triggering

FA adv FA-CoA

Co-CoA

IKE/PANA

AAA

IPSec MIP

Registrati

on

FA

RFA

HA

MIPv6 Any Stateless

Proactive

CARD

802.21

11R

Router

Adv.

Router

Prefix

CoA IKE/PANA

AAA

IPSEC MIP

update

MIP RO

CH

MAP

HA

SIPM Any Stateless

ICMP

Router

802.21

11R

L3

Router

Adv.

Router

Prefix,

ICMP

CoA

AOR

Re-Register

INVITE

exchange/AA

A

IPSEC/

SRTP/

S/MIME

Re-

INVITE

B2BUA

CH

RTPtrans

Functional Matrix of Mobility Event


Mobility

Event

Network

discovery &

selection

Network

attachment

Configuration Security

association

Binding

update

Media

reroute

Channel

discovery

L2

association

Router

solicitation

Domain

advertisement

Identifier

acquisition

Duplicate

Address

Detection

Address

ResolutionAuthentication

Key

derivation

Identifier

update

Identifier

mapping

Binding

cache

Tunneling

Buffering

Forwarding

Bi-casting/

Multicasting

Server

discovery

Identifier

Verification

Subnet

discovery

P1 P2 P3 P4 P5 P6

P11

P13

P12

P21

P22

P23

P31

P32

P33P41

P42P51

P52

P53

P54

P61 P62

P63

P64

System decomposition of handover process


Handover: Distributed operation across multiple layers

Time

L2

PoA

L3

PoA

Discovery Detection Configuration

Security

Association

p11

p12

p21

p31

p32 p42

p41Server

(Proxy,

/HA)

p22

Binding

Update

Media

Rerouting

p51p31

p32

p41 p42

p42p63

p62

p13p23

p31

p33

MN

p11 p12 p21 p22p31 p41

p61p32 p42

p13 p23p33

p51

p51

p52

p52

CN

p42p52

p61

p54

p53 p54

p61

p61p62

p64p51


Inter-domain Handoff Delay Analysis (example)

Operation

L2

Delay

L 2 Scanning

Association

L2 security

L3

Delay

Address

Acquisition

Duplicate

Address

Detection

ARP

Update

Local

Authentication

AAA

Profile

Binding

Update

Media

RedirectionApplication

Layer

Delay

-Reduce the handoff delay

-Reduce the packet Loss


An abstract view of mobility


Why need a mobility model ?Optimization techniques for a mobility event can be designed based on precedence relations amongst events and concurrent, conflicts or resource sharing type operations

Need a framework and model

– to analyze and schedule handoff processes for systems optimization

– to conduct trade-off analysis between systems resources and performance metrics

Specific expected results

– Determine the extent of parallelism possible among the handoff operations based on dependency and

– Determine systems performance (e.g., handoff delays) based on the execution of primitive handoff operations under constraints of limits on parallelism and constraints on the use of shared resources

– A mechanism to verify and predict the systems performance of a specific optimization technique

– A mechanism that can help design the optimal path of sequence of execution of events

35


Specifics of IP-mobility model Mobility event exhibits concurrent, sequential, conflicts or resource sharing behavior

Handoff-related processes can be modeled as Discrete Event Dynamic Systems (DEDS) that span across multiple layers

Proposed approach to build a mobility model

– Determine data dependency among mobility events

– Determine the consumption of shared resources the handoff operations

– Apply Deterministic Timed Transition Petri Net (DTTPN) to build various un-optimized mobility models and their associated optimization techniques

– Evaluate and predict the performance of the handoff system that demonstrates parallelism, optimistic or predictive operations

36


The system model can be used to investigate parallelism and opportunities for optimization during a handoff operation. Using the model, one can predict or verify the systems performance of an un-optimized handover and of any specific handoff optimization technique.

The model can predict the performance of any mobility protocol in any specific deployment scenario, such as intra-domain, inter-domain, or heterogeneous handoff.

The model can also be used to analyze the trade-off between performance metrics and resources when a mobility event includes parallel, optimistic, or speculative operations.

Key Benefits of Mobility Model



Description of Places and Transitions


Petri net dependency of mobility eventsHandoff Process Precedence

RelationshipData it depends on

P11 – Channel Discovery P00 Signal-to-Noise Ratio value

P12 – Subnet discovery P21,P22 Layer 2 beacon ID

L3 router advertisement

P13 – Server discovery P12 Subnet address

Default router address

P21- Layer 2 association P11 Channel number

MAC address

Authentication key

P22- Router solicitation P21, P12 Layer 2 binding

P23- Domain advertisement P13 Server configuration

Router advertisement

P31 – Identifier acquisition P23,P12 Default gateway

Subnet address

Server address

P32 – Duplicate address

Detection

P31 ARP

Router advertisement

P33 – Address resolution P32, P31 New identifier

P41 – Authentication P13 Address of authenticator

P42 – Key Derivation P41 PMK (Pairwise Master Key)

P51 – Identifier update P31,P52 L3 Address

Uniqueness of L3 address

P52 – Identifier verification P31 Completion of COTI

P53 – Identifier mapping P51 Updated MN address

at CN and HA

P54 – Binding cache P53 New Care-of-address mapping

P61 – Tunneling P51 Tunnel end-point address

Identifier address

P62 – Forwarding P51, P53 New address of the mobile

P63 – Buffering P62, P51 New identifier acquisition

P64 – Multicasting/Bicasting P51 New identifier acquisition


Handoff Transitions and Sub-transitions

Transition v Handoff Operation Sub transitions Sub-operations

t0 Disconnect trigger t00 Layer 2 un-reachability test

t01 Layer 3 unreachability

t1 Network

discovery

t11 Discover layer 2 channel

t12 Discover layer 3 subnet

t13 Discover server

t2 Network

attachment

t21 Layer 2 association

t22 Router solicitation

t23 Domain advertisement

t3

configuration

t31 Identifier acquisition

t32 Duplicate address detection

t33 Address resolution

t4 Authentication t41 Layer 2 open authentication

t42 Layer 2 EAP

t5 Security association t51 Master key derivation

t52 Session Key derivation

t6 Binding update t61 Identifier update

t62 Identifier verification

t63 Identifier mapping

t64 Binding cache

t7 Hierarchical binding

update

t71 Fast binding update

t72 Local caching

t8 Media redirection t81 Tunneling

t82 Forwarding

t83 Buffering

t9 Local data redirection t91 Local id mapping

t92 Multicasting/bicasting


Resource usage per mobility events


t1p1p0

t2 p2 t3

Disconnect

Trigger

ScanningL3 subnet

discoveryServer

discovery

2 1 2

32

Resources

discovered

p6

p3 p4 p5

(Resource: Battery power) (Resource: CPU cycles )(Resource: Bandwidth)

Handoff discovery process


t1p1p0 t2 p2 t3

p4p3 p5

3

1

Identifier

Acquisition

Duplicate

Address

Detection

Address

Resolution

12

2

Mobile

Configured

Mobile

Authenticated

(Resource: Battery Power) (Resource: CPU cycles)(Resource: Bandwidth)

p6

Handoff configuration process


Sub-process - 1(Identifier Acquisition)

Client is in

process of

getting IP address

Initial Client

Sends

Discover

Message

Server

Offers

Address

Client

Requests

Address

Server

Acknowledges

P1

P3

P4t1

t2 t3 t4

(Resource battery) (Resource Bandwidth) (Resource Processing power)

p5p4 p6

P2

Client is

checking the

address

Client

Waits for the

address


Sub-process - 2Duplicate Address Detection

Initial Client

Sends

ARP/Neighbor

Discovery

Client

confirms

the address

P1P2 P3

t1t2 t3

Client

Listen for

ARP response

(Resource PM) (Resource PB) (Resource PP)

3 3


Sub-process 3-IP Address Resolution (MAC-IP Address mapping)

Idle Send

ARP Broadcast

P1t1

(Resource PM) (Resource PB) (Resource PP)

33

2 2

Maps

IP address

And MAC

P 2

Network

Processing

ARP


t1p1p0 t2 p2

t3

Layer 2

associationRouter

SolicitationDomain

advertisement

2

Mobile

connected

p6

p3 p4 p5

Channel

available

(Resource: Battery Power) (Resource: CPU Cycles)(Resource: Bandwidth)

Handoff attachment process


t1p1p0

t2

WEP

Key

Open

AuthEAP

p3p2

p4

2

2

Mobile

Authenticated

p5

3

22

(Resource: Battery power) (Resource: CPU cycles)(Resource: Bandwidth)

Handoff authentication process


32 33

11 21 22 12 23

1341 42

31

52 51 53

54

64

61

62

63

00

Dependence graph for sequential operations


P00

t01

t11

t41

p11

p41

t13

p13

t42

p42

t21

p21

t22

p22

t12

p12

t23

p23 P52

t52 t51 P51

t53 p53

t64p64

t62

p62

t63

p63

t54 p54

p61

t31 t32 t33

p31 p32 p33

t00

Dependence graph for parallel operations


Mobility

Event

Network

discovery &

selection

Network

attachment

Configuration Security

association

Binding

update

Media

reroute

Channel

discovery

L2

association

Router

solicitation

Domain

advertisement

Identifier

acquisition

Duplicate

Address

Detection

Address

ResolutionAuthentication

Key

derivation

Identifier

update

Identifier

mapping

Binding

cache

Tunneling

Buffering

Forwarding

Bi-casting/

Multicasting

Server

discovery

Identifier

Verification

Subnet

discovery

P1 P2 P3 P4 P5 P6

P11

P13

P12

P21

P22

P23

P31

P32

P33P41

P42P51

P52

P53

P54

P61 P62

P63

P64

System decomposition of handover process



Layer 2 Optimization


802.11 Networks

A handoff occurs when a mobile station moves

beyond the radio range of one AP and enters

another BSS.


Layer 2 delay

900 ms900 ms

802.11 802.11


Layer 2 discovery process (802.11)

State1 Unauthenticated

Unassociated

State 2Authenticated

Unassociated


Associated

Successful

Authentication

Successful

Authentication or

Re-association

Disassociation

Notification

De-authentication

Notification

De-authentication

NotificationClass 1

Frames

Class 1 & 2

Frames

Class 1, 2 &3

Frames

Class 1 Frames – Control Frames

Class 2 Frames – Management Frames

Class 3 Frames – Data Frames

State1 Unauthenticated

Unassociated


Unassociated


Associated

Successful

Authentication

Successful

Authentication or

Re-association

Disassociation

Notification

De-authentication

Notification

De-authentication

NotificationClass 1

Frames

Class 1 & 2

Frames

Class 1, 2 &3

Frames

Class 1 Frames – Control Frames

Class 2 Frames – Management Frames

Class 3 Frames – Data Frames

Discovery

Scanning

Authentication Association

Beaconing

MN

L2

PoA

MN L2

PoAMN L2

PoA

Discovery

Scanning

Authentication Association

Beaconing

MN

L2

PoA

MN L2

PoAMN L2

PoA


Layer 2 Handoff Delay (802.11)

Discovery Phase

– Active scanning

MN probes AP

– Passive scanning

AP sends beacons periodically

Authentication Phase

– Open authentication

– Shared authentication

– 802.11i – 4 way handshake

Association Phase


Layer 2 handoff sequence

Old AP

Target APsMobile

Node

Existing

association

Probe request

Probe response

Probe request

Probe response

Authentication request

Authentication response

Re-association request

Re-association response

IAPP: Send security block

IAPP: Ack security block

IAPP: Move request

IAPP: Move response

New

association

EAPOL key

EAPOL key

EAPOL key

Scanning

Delay

Authentication

Delay

4-way

handshake

delay

Association

delay

Optional


Key principles for discovery optimizationLimiting the number of signaling exchanges between the mobile and the centralized server needed to discover the network resources.

In the case of passive scanning, an increase in the rate of beacon advertisement reduces the time to discover the new point of attachment at the cost of additional network bandwidth and processing at the end hosts.

Caching of neighboring network resource parameters before the mobile moves to the new network.

Use of a media-independent application layer discovery protocol to discover network resources to support handover in heterogeneous access networks without depending upon any access-specific technology.


Layer 2 Discovery Optimization

General techniques: Reduce the scanning time Caching of ESSID Use of second interface 802.11 specific discovery Proactive Discovery (no

scanning)

Proposed Solutions: Shin et al introduces selective

scanning and caching strategy Montavont et al propose

periodic scanning Velayos et al propose reduction

of beacon interval and performs search in parallel with data transmission

Brik et al propose to use a second interface to scan while communicating with the first interface

802.11u, 802.11k Forte and Schulzrinne Application Layer proactive

discovery (e.g., Dutta et al)


Expreriment Result – Handoff

time

Handoff Time

0

100

200

300

400

500

600

1 2 3 4 5 6 7 8 9 10

Experiments

mse

cOriginal HandoffSelective ScanningCaching


Optimization of Layer 3 Configuration


Components that affect L3 configuration and optimization techniques for layer 3 configuration

Layer 3 address acquisition

– Proactive caching

Duplicate Address Detection

– Optimistic DAD, Proactive DAD, Passive DAD,

– Router Assisted DAD

NUD (Neighbor Unreachability Detection)

– Aggressive Router Selection

Configuration

Identifier

AcquisitionDuplicate

Address

Verification

Identifier

Mapping

Layer 2

Layer 3

Mobile

NodeServer Network

Mobile

Node L3 POA Network

MNServer

L3

PoA

Configuration

Identifier

AcquisitionDuplicate

Address

Verification

Identifier

Mapping

Layer 2

Layer 3

Mobile

NodeServer Network

Mobile

Node L3 POA Network

MNServer

L3

PoA


Key principles for Layer 3 Configuration

Reduction of the number of signaling messages exchanged between the mobile node and the DHCP server during stateful IP address acquisition.

Minimizing the time taken to verify the uniqueness of the IP address of the mobile.

Performing the address uniqueness checking ahead of layer 3 handoff.

Prefetching and caching of the new IP address reduces the time taken for IP address acquisition after the handoff.

Performing the address resolution by mapping between the IP address of the target router and the MAC address before the mobile has moved to the new network.


MIPv6 and SIP-Mobility experimental Testbed


Experimental Results – IPv6


Authentication Optimization


How security related protocols affect performance

Security protocols have an impact on the performances of the network

– End-to-end latency

– Throughput

– Handoff delay

Main components that affect the performance

– Authentication/authorization, Key Derivation, Encryption

Security related delays may affect all the layers

Layer 2 (e.g., 802.11i, WEP)

Layer 3 (IPSEC/IKE)

Upper Layers (e.g., TLS, SRTP)

Security

Association

Key

Distribution Authentication Encryption

Layer 2

Layer 3

Layer 4

ServerMobile Network

MN

MN Server

L3

POA

Security

Association

Key

Distribution Authentication Encryption

Layer 2

Layer 3

Layer 4

ServerMobile Network

MN

MN Server

L3

POA


Key principles for Authentication optimization

Minimization of the time needed to authenticate and authorize the mobile after each handoff during the re-authentication procedure.

Reduction of the number of signaling messages that need to be exchanged between the mobile node and the authenticator to generate a shared secret key.

Use of an appropriate key generation algorithm that reduces the processing load on the end hosts.

Placement of the authenticator and authentication server closer to the mobile.

Reduction in installation time of the pre-shared keys (PSKs) on the authenticator in the case of proactive authentication.

Proactive caching of the security context at the neighboring access points prior to handoff, either by proactive authentication or by context transfer.


Optimizing authentication Related Work

IEEE Standards– IEEE 802.11i provides pre-authentication at link-layer in the

distribution system (DS)– IEEE 802.11r improves 11i by introducing a new key

hierarchy but it does not work between DSs either.

Context transfer solutions (Bargh et al, Georgiades et al, Duong et al)– Security problems such as “domino effect”– Assume certain trust relationships which might not be

possible in certain scenarios.– Oriented towards the same technology

Re-authentication

Pre-installation based on movement pattern (Mishra et al, Pack et al )– AAA assisted key installation– Works within the same administrative domain

MIPv6 and AAA assisted (Ruckforth et al)– Limited to MIPv6 and within the same domain

Cooperative Roaming (Forte et al)– Works within a domain


Authentication Optimization

Authentication mechanism requires 802.1x message exchange with the authenticator in the target network

Number of round trip signaling and key derivation process need to be minimized

Low latency re-authentication

Authentication can be done proactively

Context can be transferred

Layer 3 authentication bootstraps layer 2 authentication process


Network-Layer Assisted Pre-Authentication Technique

Assists link-layer optimization mechanism to work accross subnets and domains

It is independent of link-layer technology (e.g., 802.11, CDMA)

It does not suffer from context transfer security problems and only assumes basic trust relationship

It supports handover across inter-technology, inter-subnet and inter-domain.


Experimental Testbed

Home AAA

Domain

IEEE 802.11i

Pre-authentication

nAR/PAA

AAAv

AAAh

pAR165.254.55.116/24

165.254.55.115/24

155.54.204.82

10.1.30.1/24

10.1.30.3/2410.1.30.2/24

10.1.10.2/24

10.1.10.1/2410.1.20.2/2410.1.20.1/24

MN

PSK PSK

AP0AP1AP2

Radius/Diameter

PANA pre-auth

Association

&

4-way handshake

Network A Network B

PANA Pre-authentication

Roaming AAA

Domain*

* Roaming AAA Domain in roaming case.

For non-roaming case, it acts as MN’s home AAA

domain.

Non-Roaming: [email protected]

Roaming: [email protected]




Results (II)


Optimizing a) Binding Updateb) Media Rerouting


Optimizing Binding Update

Techniques– Reduce the latency due to

longer binding update when the communicating host is far away

– Limit the binding update within a domain

Proposed Solutions– IDMP– Regional registration-

based Mobile IP– HMIPv6– Anchor-based Application

Layer B2BUA

– Proactive Binding Update

Binding

Update

Tunneling Mapping Caching

Mobile Network Anchor Mobile CN

Anchor

PointCN

Binding

Update

Tunneling Mapping Caching

Mobile Network AnchorMobile Network Anchor Mobile CN

Anchor

PointCN


Key principles for Layer 3 Binding UpdateLimiting the traversal of the binding update closer to the mobile after every handoff.

Use of two levels of binding update by using an anchor agent between the home agent and the mobile node.

Applying the binding update proactively in the previous network before the mobile has moved to the new network.

Simulcasting the data to help reduce the data loss due to a longer binding update delay. This can probably be achieved by using a localized multicast approach.


Media redirection and optimized binding update for SIP-based mobility

Capture the transient packets in-flight and redirects to the mobile–SIP Registrar and NAT-like functionality

RTPtrans (RTP translator an application layer Translator)

Mobility Proxy (Linux iptables)

–Outbound SIP proxy server

Local SIP proxy captures outbound packets

Limit the signaling due to Intra-domain Mobility–B2B SIP UA

Emulates Third Party Call control

–Multicast Agent

–Small group multicast

–Duration limited locally scoped Multicast


SIP-based Fast-Handoff

MN

Internet

Visited Domain

MN

MN

Public SIP Proxy

Public SIP Proxy

Public SIP Proxy

IP0

IP1

IP2

Visited

Proxy

Home SIP

Proxy

RTP

Media

(Existing SIP

Session)

OKACK

CNHome

Domain

Subnet

S0

Subnet

S1

Subnet

S2

RTP

Media after

Re-Invite

Register

1

2

3

4

5

Translator

Translator

Translator


Hierarchical Mobility Management IDMP+MIP

Home Network

1

2

1

3

2

MA

SA

MN

• All packets from the global Internet tunneled (re-directed) to the

GCoA and are intercepted by the MA.

• MA tunnels each packet to the MN’s current LCoA.

CN

SASA

HADomain


Experimental Results on Mobility Optimization(Systems Evaluation)


Experimental Validation of Mobility OptimizationCase Studies

Following are the experimental case studies where we have beenable to optimize the handoff delay and reduce the packet loss bydeploying several Optimization Techniques

Case I - Optimizing data path between CH and MH Case II - Optimizing Binding Update Case III - Optimizing Layer 3Case IV - Optimizing Security AssociationCase V - Make-before-Break Technique Case VI - Maintaining Security Association Case VII - Media Independent Pre-authentication proactive handover and bufferingCase VIII – Optimized IMS HandoffCase IX - Multicast Mobility


Media-independent Pre-Authentication

MPA is:

–a mobile-assisted higher-layer authentication, authorization and handover scheme that is performed a-priori to establishing L2 connectivity to a network where mobile may move in near future

MPA provides a secure and seamless mobility optimization that works for

–Inter-subnet handoff

–Inter-domain handoff

–Inter-technology handoff

Use of multiple interfaces

MPA works with any mobility management protocol


Functional Components of Proactive Handoff

1) Pre-authentication/authorization

– Used for establishing a security association (SA) between the mobile and a network to which the mobile may move

2) Pre-configuration

– Used for obtaining parameters (e.g., an IP address) from the network to which the mobile may move

– The SA created in (1) are used to perform secured configuration procedure

3) Secured Proactive Handover (PH)

– Used for sending/receiving IP packets from the current network using the pre-configured parameters of the new network


Media-independent Pre-Authentication (MPA)

MPA is a mobile-assisted higher-layer authentication, authorization and handover scheme that is performed a-priori to establishing L2 connectivity to a network where mobile may move in near future

MPA provides a secure and seamless mobility optimization that works for Inter-subnet handoff, Inter-domain handoff and Inter-technology handoff

MPA works with any mobility management protocol

TimeConventional

Method

AP DiscoveryAP

Switching

MPA

Pre-authentication

IP address

configuration

& IP handover

Time

Client

Authentic

ation

Packet Loss Period


Media Independent Pre-authentication -

Seamless Handoff (a deployment scenario)

AA CA

MN-CA keyAR

Network 3

AR

AA CA

MN-CA key

Network 2

INTERNET

Information

Server

Mobile

Current

Network 1AR

AP1 Coverage Area AP 2 & 3 Coverage Area

AR

Network 4

CN

AP3AP2AP1 CTN

TN

CTN – Candidate Target Networks

TN – Target Network


Home

Network HA

MPA Overview

CN: Correspondent Node

MN: Mobile Node

AA: Authentication Agent

CA: Configuration Agent

AR: Access Router

AA CA

A(X)

2. DATA [CN<->A(Y)]

over proactive handover

tunnel [AR<->A(X)]

AR

L2 handoff

procedure

Domain X Domain Y

CN

Data in new

domain

1. DATA[CN<->A(X)]

MN-CA key

Pre

configuration

pre-authentication

MN-AR key

3. DATA[CN<->A(Y)]

Data in old

domain

MN

A(Y)

BU

Proactive handover

tunneling end

procedure

Tunneled Data

MN


Proactive Handoff Experimental Results (Case III)

Mobility Type MIPv6

Handoff

Parameters

Buffering

Disabled

+ RO

Disabled

Buffering

Enabled

+ RO

Disabled

Buffering

Disabled

+ RO

Enabled

Buffering

Enabled

+ RO

Enabled

Buffering

Disabled

Buffering

Enabled

L2 handoff

(ms)

4.00 4.33 4.00 4.00 4.00 5.00

Avg. packet

loss

1.33 0 0.66 0 1.50 0

Avg. inter-

packet interval

(ms)

16.00 16.00 16.00 16.00 16.00 16.00

Avg. inter-

packet arrival

time during

handover (ms)

n/a 45.33 n/a 66.60 n/a 29.00

Avg. packet

jitter (ms)

n/a 29.33 n/a 50.60 n/a 13.00

Buffering

period (ms)

n/a 50.00 n/a 50.00 n/a 20.00

Avg. Buffered

Packets

n/a 2.00 n/a 3.00 n/a 3.00

SIP Mobility


Performance (MPA-Non-MPA) – Single I/F

MPA– No packet loss during pre-

authentication, pre-configuration and pro-active handoff before L2 handoff

– Only 0 packet loss, 4 ms delay during handoff mostly transient data Includes delay due to layer 2,

update to delete the tunnel on the router

We also reduced the layer 2 delay in hostap

Driver L2 delay depends upon driver

and chipset

non-MPA– About 200 packets loss, ~ 4 s

during handover Includes standard delay due to

layer 2, IP address acquisition, Re-Invite, Authentication/Authorization

– Could be more if we have firewalls also set up

MPA Approach

Non-MPA Approach

handoff

802.11 802.11

4 s


Handoff Delay

~ 18 s

802.11 CDMA

Handoff Delay

16 s

802.11 CDMA

a. MIP-based Non-optimized handoff

b. SIP-based Non-optimized handoff

c. MPA and 802.21 assisted optimized

handoff

802.11 CDMA

Optimized handoff delay with MPA (Multiple I/F)


Optimization in IMS Testbed

P-CSCFP-CSCF S-CSCF

AS

HSS

I-CSCF

PDSN HA

VN1-re2VN2-re3

802.11b 802.11b

Visited Network 1

Visited Network 2

DHCPDHCP

RAN Emulator

Mobile Node

K6Router

192.168.6.0/24192.168.8.0/24

6.2

6.1

8.2HN-HA

HN-AS-SCSCFHN-HSS-ICSCF

VN1-PCSCF

VN1-DHCP

VN2-PCSCF

VN2-DHCP

VN2-PDSN

VN1-PDSN

VN1-RE-12

PDSN

RAN Emulator

VN2-RE-21

8.1

Mobile Node

Domain: kddi.testbed

VN1-re1

802.11b

Home Network

IPTV Server

HN-IPTVServer

RAN Emulator

VN1-RE-11

6.3

PDIF

VN2-PDIF

VN2-re4

802.11b

::5::10::15::25::5::10::15

3ffe:2::/64

3ffe:1::/643ffe:5::/64

::1

::1

::1

::10::5

3ffe:5::30

(Mobile IP case) mh2

3ffe:5::35

(Mobile IP case)

::20::15::25

PDIF

VN1-PDIF

VN1-re5

802.11b

<PPP address on PDSN>

mh1 3ffe:11::MAC/64

mh2 3ffe:11::MAC/64

<PPP address on PDSN>

mh1 3ffe:22::MAC/64

<Address on PDIF>

mh1 3ffe:33::MAC/64

<Address on PDIF>

mh1 3ffe:44::MAC/64

PCRF

VN2-PCRF

::30

VN1-PCRF

::30

PCRF

To visited domain

mh3

3ffe:5::40

(Mobile IP case)

Mobile NodeMobile Node3ffe:5::30

(Mobile IP case)

Current demonstration

• P-CSCF fast handover– Non-Optimized

– Reactive

– Proactive

• Optimized Roaming– Dual anchoring

– Home address anonymity


0

1458

1502

0

1501

1408

0

2399

5980

89

200

195

0 3000 6000 9000 12000

Proactive

Reactive

Non-Optimized

Time in ms

Typ

es o

f H

an

do

ff

Link (PPP) Termination

Layer 2 (802.11) Delay

Link (PPP) Activation

MIP-Solicitation

MIP-Binding Update

DHCP Trigger

DHCP Inform

SIP Registration

SIP(AKA) Security

Media Redirection

Handoff components optimized


Layer 2 handoff Model


Petrinet Model for Layer 2 Handoff – Sequential


Petrinet Model for Layer 2 Handoff – Concurrent


Scheduling of handoff operations

97

Association

Network

discovery

P11

t11

PA2

4-way

Handshake

(SA)

t1

t4 t5

P2 P3

Connected

Dis

connected

Pre-

authentication

Current Network Target Network

PA1

PC

PB1

PD

t12

t13

AP

Key

installation

P12

P1

Resources CPUPC

Resource s BatteryPB

4-way handshakecompletet3

t4 t5

P2

P3

t2

Scanning

Authentication

NetworkDiscovered

4-wayHandshakeOperation

P1

ResourcesNetwork capacity

MobileAuthenticated

Connected

Association

P0

P01

P02

2 2

t1

PA

PC CPU

BatteryPB

t3

t4

t5

P2

t2

Scanning

Authentication

NetworkDiscovered

4-wayHandshake

P1

ResourcesNetwork Capacity

MobileAuthenticated

Connected

P0

P01

P02

2

t1

P03

P3Association

4

PA

C. Proactive operations

B. Parallel operations – Level of concurrency =2

D. Parallel operations – Level of concurrency = 3

A. Sequential operations

Battery

power

scanning Authentication 4-way

Handshake

t2 t3 t4 t5

P2 P3 P4

Association

Connected

Mobile

Disconnected

Network

capacity

CPU

cycles

P1

PA

PB

PC

P0

t1Disconnection

Network

Discovered

Mobile

authenticated

1 token


Deadlock analysis for simultaneous mobility using MATLAB models

98

Deadlock Scenario (non-optimized) Deadlock verification (deadlock exists)

Deadlock avoidance with retransmission Deadlock verification (No deadlock)


Summary of Experimental results for optimization techniquesHandoff components Optimization techniques

Discovery Application layer proactive discovery

Authentication Network layer assisted layer 2 pre-authentication

Layer 3 security association Anchor assisted security association

Proactive security context transfer

Layer 3 configuration Router assisted duplicate address detection

Proactive IP address configuration

Route optimization Maintain direct path

Interceptor assisted packet modifier

Intercepting proxy assisted route optimization

Binding cache-based route optimization

Binding update Hierarchical binding update

Proactive binding update

Proactive proxy-based join for multicast traffic

Simultaneous mobility

Media rerouting Data redirection using forwarding agent

Mobility proxy assisted time-bound data redirection

Time bound localized multicasting

Media buffering Dynamic buffer control protocol

Cross layer triggers Media independent handover primitives


100

Scheduling

types

Relevant

optimization

principles

Example experimental mobility systems Potential

Target

Mobility

System

SIP-based

Fast

handoff

Mobile

VPN

Media

Independent

Pre-authentication

Simultaneous

Mobility

Optimized

handoff

In IMS

Muti-layer

Mobility

Multicast

fast

handoff

Sequential Direct path between

CH and MHX

Limit binding update

between CH and MHX X

Maintain Security

association

between end-points

X √

Anchor-based

ForwardingX X √

Post-handoff triggers X

Predictive Pre-handoff triggers X X

Proactive network

discoveryX

Proactive

authentication X

Proactive identifier

configurationX √

Proactive

binding updateX X

Dynamic Buffering X

Proactive context

transferX

Parallel Discovery of Layer 2

and

Layer 3 PoA

X √

Binding update

during configurationX

Target mobility system design


Key Takeaway• Identification of fundamental properties that are rebound during a

mobility event. Analysis of these properties provides a systematic framework for describing mobility management and the operations that are intrinsic to handover.

• A model of the handover process that allows one to predict performance both for an unoptimized handover and for specific optimization methodologies under conditions of resource constraints. This model also allows one to study behavioral properties of the handoff system such as data dependency and deadlocks.

• A series of optimization methodologies, experimental evaluations of them, and optimization techniques that can be applied to the link, network, and application layers and preserve the user experience by optimizing a handover.

• Application of the model to represent optimizations, and comparison of the results with experimental data.


Since the current mobility protocols and associated optimization techniques are ad hoc in nature, it is useful to have a systematic analysis of the mobility event when designing appropriate optimization techniques.

Since mobility involves various layers of the protocol stack, it is important to discover the type of mobility that a mobile will be subject to, such as layer 2, layer 3, or application layer mobility.

– The type of mobility is determined by the mobile node’s mobility pattern, such as cell handoff, subnet handoff, or domain handoff, the type of application supported on the mobile node, and the type of access network.

Since layer 2 handoff optimization techniques are access-dependent, it is important to consider the access characteristics of each network, such as the channel access algorithm (e.g., CSMA/CA, OFDM, or TDMA). For example, a CDMA network will have different access characteristics from an 802.11 network. The amount of resources used (e.g., channel bandwidth) will vary with the type of access network.

Mobility Optimizaion Best Current Practice


Each mobility event (e.g., handoff) can be considered to consist of a set of primitive functions, such as discovery, configuration, authentication, security association, registration, binding update, and media delivery. Optimizations of these primitive functions can take place independently of each other but often benefit from cross-layer triggers.

A mobility event can be considered as a discrete-event dynamic system, where each of the abstract functions can be considered as a specific discrete event. Optimizing each of the discrete events can contribute to the overall optimization of the system.

The scheduling of the primitive functions that are part of these handoff events plays an important role in the overall systems behavior, including systems performance and resource usage.

The scheduling of the handoff primitives needs to take account of the data dependency among the abstract operations. The data dependency will determine the extent of parallelism that is possible during the handoff operations.

Best Current Practice – contd.


Deadlocks need to be avoided during any mobility operation. Deadlocks are typically caused by a lack of data from previous primitive operations or a lack of resources needed for an operation.

Thus, the scheduling of the primitive events should ensure that there are enough resources available for parallel or speculative operations of any kind and that data is available.

It is important to consider the type of transport (e.g., RTP or TCP) supported by an application running on the mobile when it is subjected to handoff, as each of these applications has different performance requirements in terms of packet loss, delay, and jitter.

Since there are several mobility protocols available and each of these protocols is suitable for a specific type of application (e.g, RTP- or TCP-based transport) and a specific type of handoff (e.g., layer 2, layer 3, or interdomain handoff), a policy-based mobility management scheme can be appropriate in many cases.



Since the primitive handoff operations in each layer take place independently of the operations in other layers, cross-layer triggers from lower layers can help to expedite the handoff operations in the upper layers. Thus, any optimization framework needs to apply some of the available cross-layer optimization techniques. IEEE 802.21 has defined a Media Independent Handover Function that provides cross-layer triggers to expedite a handover.

It is always useful to have a handoff model that can predict the systems performance based on the schedule and the available systems resources. When the systems parameters and resource availability are varied, the performance of the system will also vary. Service providers can use such a handoff model to determine what types of protocol and optimization techniques are needed in a specific scenario.

The scheduling of handoff primitives is largely determined by the systems resources and the data dependency among the events. Since the scheduling of handoff primitives affects the systems performance, it can be changed to meet performance requirements at the cost of added systems resources.



The scheduling of handoff operations can also affect the trade-off between the resources expended (e.g., battery, CPU, and bandwidth) and systems performance (e.g., delay and packet loss). Thus, the types of optimization that should be used are largely determined by the extent of the trade-off that can be allowed against resources.

In the case of multi-interface mobility, a make-before-break mechanism helps to reduce the delay and packet loss at the cost of additional resources,1 since both of the interfaces remain active during handoff. The extent of overlap of the operations is determined by the amount of resources that can be expended during handoff.

Proactive operations appear to be more attractive for providing the desired handoff performance (e.g., delay and packet loss) compared with sequential and parallel operations. However, there is a trade-off between the amount of resources and the performance when there are multiple target networks, since the mobile needs to complete proactive handoff-related operations with multiple target networks to increase the probability of a successful handover.



The mobility model could be enhanced to study the behavioral properties and systems performance of any type of mobility protocol, such as transport layer protocols and mobility in other types of networks such as ad hoc networks.

The model could be enhanced so that one could use it in an automated fashion to generate a specific schedule for the handoff operations given a set of resource constraints and performance objectives, and a dependency graph. Automatic generation of schedules for handoff operations to provide the desired quality of service with the available resources will help one to use the right set of protocols.

Using a systematic analysis of the mobility functions, one can design a customized mobility protocol suitable for one’s own set of requirements.

This model could be enhanced to predict performance based on the resource parameters of all of the network elements that are involved in the mobility event.

The formalization of key techniques, the models of systems dependencies, and the ability to calculate or predict optimization metrics provide a foundation for the automated discovery and implementation of mobility optimization.

Future work for Research


"Mobility Protocols and Handover Optimization: Design, Evaluation and Application" written by Ashutosh Dutta and Henning Schulzrinne and published by Wiley-IEEE Press in 2014. (ISBN 978-0-470-74058-3, Hardcover, 476 pages.)

Save 20% when you enter code VBG91.

Reference Book

http://www.wiley.com/WileyCDA/WileyTitle/productCd-0470740582.html

Documents

Mobility Protocols and Handover Optimizationsites.ieee.org/denver-com/files/2018/05/Boulder-DL-Talk.pdf2018-05-06 · 20 Types of 5G Applications Enhanced Mobile Broadband - Mobile