A Software Defined Multi-Path Traffic Offloading System for - - PowerPoint PPT Presentation

▶
a software defined multi path traffic
SMART_READER_LITE
LIVE PREVIEW

A Software Defined Multi-Path Traffic Offloading System for - - PowerPoint PPT Presentation

Apr 19, 2023 108 likes •410 views

UCLA ENGINEERING Computer Science A Software Defined Multi-Path Traffic Offloading System for Heterogeneous LTE-WiFi Networks Pengyuan Du * , Qi Zhao * and Mario Gerla Department of Computer Science, UCLA IEEE WoWMoM 2019 6/11/2019 *

slide-1
SLIDE 1

UCLA ENGINEERING Computer Science

A Software Defined Multi-Path Traffic Offloading System for Heterogeneous LTE-WiFi Networks

Pengyuan Du*, Qi Zhao* and Mario Gerla

Department of Computer Science, UCLA IEEE WoWMoM 2019 6/11/2019

* Co-primary author

slide-2
SLIDE 2

UCLA ENGINEERING Computer Science

In Memory of Professor Mario Gerla

❑“The imprint Prof. Gerla made on all our lives is indelible. He was a beloved husband, father, brother to his family, a thought leader and brilliant mind within the global computer science community, a mentor and inspiration to his Masters and PhD students, and a dear friend to us

  • all. May we all

honor his spir- it by living life with passion, dedication and

  • pen minds.”

75th Birthday

slide-3
SLIDE 3

UCLA ENGINEERING Computer Science

Outline

❑ Introduction ❑ Software-Defined Multi-Path Traffic OPtimization (SD-MPTOP)

➢ Traffic Offloading in HetNet ➢ SD-MPTOP Overview ➢ Multi-Commodity Flow Modeling

❑ SD-MPTOP Implementation ❑ Performance Evaluation ❑ Conclusion

slide-4
SLIDE 4

UCLA ENGINEERING Computer Science

Introduction

❑The era of mobile communications

Mobile phone Vehicle Home

slide-5
SLIDE 5

UCLA ENGINEERING Computer Science

Introduction

❑The explosion of mobile communications[1]

➢ Sevenfold increase in mobile data traffic ➢ Spectrum crisis

[1] Cisco. “Cisco Visual Networking Index: Forecast and Methodology, 2016–2021.”, 2018. 10 20 30 40 50 60 2016 2017 2018 2019 2020 2021

Mobile data traffic (Exabyte)

slide-6
SLIDE 6

UCLA ENGINEERING Computer Science

Introduction

❑ Offloading cellular traffic

➢ Use of alternative network technologies for data delivery ➢ Infrastructure

  • Small-cell / WiFi / 3GPP

➢ Opportunistic networks ( aka Mobile ad-hoc networks)

  • Delay tolerant networks (DTN)
  • Device-to-Device (D2D) communications

Micro cell Femto cell Macro cell Pico cell

Multi-home

DTN

D2D

Content distribution

slide-7
SLIDE 7

UCLA ENGINEERING Computer Science

Introduction

❑Heterogeneous infrastructure-based network (HetNet)

➢ Current practice: direct use of WiFi ➢ Our solution: multipath, concurrent use of cellular and WiFi

  • User: bandwidth aggregation
  • Network operator: centralized management
slide-8
SLIDE 8

UCLA ENGINEERING Computer Science

Traffic Offloading in HetNet

Internet

LTE

Internet

LTE WiFi

Internet

LTE WiFi

Internet

LTE

Internet

LTE WiFi LTE

MPTCP

slide-9
SLIDE 9

UCLA ENGINEERING Computer Science

Multipath TCP Protocol

❑Multipath TCP (MPTCP)[1]

➢ Presenting a single TCP connection to the application ➢ Utilize different interfaces underneath ➢ Work over today’s networks

[1] Wischik, Damon, et al. "Design, Implementation and Evaluation of Congestion Control for Multipath TCP." NSDI, 2011.

slide-10
SLIDE 10

UCLA ENGINEERING Computer Science

Internet

HetNet 3 WiFi HetNet 1

Current practice single path user selection

Traffic Offloading in HetNet

❑Resource allocation problem

Internet

HetNet 1 HetNet 2 HetNet 3

Server HetNets Devices Physical connectivity Capacity

Internet

HetNet 2 HetNet 1 HetNet 3

Multipath

Internet

WiFi HetNet 1 HetNet 3

Global view

  • ptimal selection
slide-11
SLIDE 11

UCLA ENGINEERING Computer Science

Software Defined Networking

❑SDN architecture

➢ Decoupling of control and data plane

  • Centralized management
  • Network function virtualization

Network Operating System Routing

Traffic Engineering Other Applications

Well-defined API Network Map Abstraction Forwarding Forwarding Forwarding Forwarding Separation

  • f

Data and Control Plane Network Virtualization

Security

D a D at a t a P P l a l ane ne C ont rol P l a l ane ne Application Plane

Instructions Instructions Instructions Instructions

slide-12
SLIDE 12

UCLA ENGINEERING Computer Science

Traffic Offloading in HetNet

❑Design considerations

i. Efficient bandwidth aggregation ii. Service continuity

  • iii. Computation hardness
  • iv. Network management, QoS, Deployment cost

❑Proposed solution architecture

➢ Multipath TCP (MPTCP) ➢ Software Defined Networking (SDN)

MPT PTCP SD SDN

slide-13
SLIDE 13

UCLA ENGINEERING Computer Science

SDN Switch

System Overview

MPTCP

WiFi

subflow1 subflow2 subflow1 subflow2

Controller MP Assoc.

MPTCP SYN/ACK 1 1 0 0 0 1 1 1 1

MCF Solver

Network report

Config

Data plane Control plane

slide-14
SLIDE 14

UCLA ENGINEERING Computer Science

Traffic Offloading in HetNet

Internet WiFi

. . .

Controller

MPTCP MPTCP Subflow1 Subflow1 Subflow1 Subflow1

MP Assoc.

WiFi LTE

MN1 MN2 MNk

. . . . . .

0 1 1 1 1 0

. . . . . .

WiFi LTE

MN1 MN2 MNk

. . .

0 1 1 1 1 0

. . .

  • Conn. Matrix
slide-15
SLIDE 15

UCLA ENGINEERING Computer Science

Traffic Offloading in HetNet

Internet WiFi

. . .

Controller

MPTCP MPTCP

RSSI Net Load Capacity RSSI Net Load Capacity

MCF Solver

MP Assoc.

WiFi LTE

MN1 MN2 MNk

. . . . . .

0 1 1 1 1 0

. . . . . .

WiFi LTE

MN1 MN2 MNk

. . .

0 1 1 1 1 0

. . .

  • Conn. Matrix

WiFi LTE

MN1 MN1 MNk

Bandwidth Allocaion

slide-16
SLIDE 16

UCLA ENGINEERING Computer Science

SDN Switch SDN Switch

Traffic Offloading in HetNet

Internet WiFi

. . .

Controller

MPTCP MPTCP MCF Solver

MP Assoc. Config

Switch Configurations Flows Src_ip: 10.0.3.0 Src_ip:10.0.3.1 Src_ip:10.0.2.0 Src_ip:10.0.2.1 Action Queue1,Output:3 Queue2,Output:4 Queue3,Output:3 Queue4,Output:4

WiFi LTE

MN1 MN1 MNk

Bandwidth Allocaion

slide-17
SLIDE 17

UCLA ENGINEERING Computer Science

Traffic Offloading in HetNet

Software-Defined MultiPath Traffic OPtimization (SD-MPTOP)

Internet WiFi

. . .

Controller

MPTCP MPTCP MCF Solver

MP Assoc. Config

slide-18
SLIDE 18

UCLA ENGINEERING Computer Science

SD-MPTOP MCF

❑Multi-Commodity Flow (MCF)

Internet

1

BW

2

BW

3

BW

1

R

2

R

3

R

4

R Net1 Net2 Net3 MN1 MN2 MN3 MN4 Subject to:

Capacity constraint Requirement constraint

min Obj( )

1,1

f

1,2

f

3,2

f

3,3

f

, n s

f

, n s

f

, n s s n

f BW 

, n s n s

f R =

slide-19
SLIDE 19

UCLA ENGINEERING Computer Science

SD-MPTOP MCF (cont’d)

❑Routing Assignment

➢ Objective function: minimizes the average packet delay

,

, , , ,

1 min

s n s s s

n s n N n s n s s s n N s n s n N link l d f

  • a

f T f d BW f 

    

      = +    −      

    

Service rate

Flow rate

,

s

n s n N

f

s

BW

[1] Fratta, Luigi, Mario Gerla, and Leonard Kleinrock. "The flow deviation method: An approach to store‐and‐forward communication network design."

Propagation delay

  • r network cost

M/M/1 Queue delay model Network congestion indicator [1] Load balancing

s

BW

,

s

n s n N

f

Load

, n s

f

Link cost

where

slide-20
SLIDE 20

UCLA ENGINEERING Computer Science

SD-MPTOP MCF (cont’d)

❑Example

,

, , , ,

1 min

s n s s s

n s n N n s n s s s n N s n s n N link load f

L f T f d BW f 

    

       = +    −      

    

WiFi LTE Bandwidth Allocation of SD-MTOP

0.06

0.50 3.50

0.44

1.00 3M 6M

MN1 MN2 MN3

LTE Bandwidth Allocation of SPTCP

3M 6M

MN1 MN2 MN3

WiFi

<1.00 <4.00 <0.50

WiFi LTE Bandwidth Allocation of MPTCP

3M 6M

MN1 MN2 MN3

<1.00 y x 4-y 0.5-x

Internet

1

R

2

R

3

R LTE WiFi MN1 MN2 MN3 =0.5Mbps =4Mbps =1Mbps

1

BW =6Mbps

2

BW =3Mbps Objective function:

  • 1. Multipath
  • 2. Load balanced
  • 1. Single path
  • 2. LTE underutilized
  • 3. WiFi congested
  • 1. Multipath
  • 2. WiFi congested
slide-21
SLIDE 21

UCLA ENGINEERING Computer Science

SD-MPTOP MCF Solution

❑ The Flow Deviation Method (FDM) ❑ Minimum allocation interruption ❑ Quick start

➢ Use previous allocation as the initial feasible solution

❑ Max-min fairness

➢ MCF could be infeasible ➢ Utilizing all the bandwidth in a fair manner ➢ Binary search the maximum demand that nearly saturates the capacity

,

, 2 , , , , , int

1 min

s n s s s s

n s n N n s n s n s n s s s n N s n N s n s n N link errruption lo d f a

f T f d f f BW f  

      

      = +  + −   −      

      

slide-22
SLIDE 22

UCLA ENGINEERING Computer Science

SD-MPTOP Implementation

WiFi

Linux TC eth0 wmediumd wlan0

MPTCP

Linux kernel MPTCP eth0 wlan0 Mobility support

Controller MP Assoc.

MCF Solver

Config

Controller Controller

draco.cs.ucla.edu

slide-23
SLIDE 23

UCLA ENGINEERING Computer Science

SD-MPTOP Evaluation

Static scenario Handover scenario Mobile scenario

slide-24
SLIDE 24

UCLA ENGINEERING Computer Science

Static scenario

Parameters

N1

5

N2

5

BW wifi

4Mbps

BW LTE

6Mbps

Delay wifi

10ms

Delay LTE

50ms

slide-25
SLIDE 25

UCLA ENGINEERING Computer Science

Handover scenario

Parameters Static, mobile 1,1 BW wifi 6Mbps BW LTE 8Mbps Delay wifi 10ms Delay LTE 50ms Req 3Mbps

slide-26
SLIDE 26

UCLA ENGINEERING Computer Science

Mobile scenario

Parameters Static 5 mobile 5 BW wifi 6Mbps BW LTE 8Mbps Delay wifi 10ms Delay LTE 50ms Req 2Mbps

Static Mobile

slide-27
SLIDE 27

UCLA ENGINEERING Computer Science

❑We designed and implemented SD-MPTOP, a centralized multi-path traffic offloading solution for HetNet

➢ Bandwidth aggregation as a resource allocation problem ➢ Centralized control, mobility management, policy ➢ Evaluated in Mininet emulation testbed

Bandwidth Aggregation Complexity Mobility Network Management Cost SPTCP

High Low

MPTCP

Low Low

3GPP

Low High

SD-MPTOP

Low Low

✘ ✔ ✘ ✘ ✔ ✘ ✔ ✔ ✔ ✔ ✔ ✔

Conclusion

slide-28
SLIDE 28

UCLA ENGINEERING Computer Science