joint scheduling and stochastic streaming for device-to

1
Joint Scheduling and Stochastic Streaming for Device-to-Device Video Delivery Joongheon Kim (USC), Andrea Turci (U. Bologna), Giuseppe Caire (USC), Andreas F. Molisch (USC) E-mails: [email protected] , [email protected] , [email protected] , [email protected] Introduction IEEE Information Theory and Applications Workshop (ITA) Graduation Day Talk (February 12, 2014) University of California, San Diego (UCSD), La Jolla, California, USA Streaming with Stochastic Network Optimization Scheduling with MWIS Formulation Transmitter Q i (t) μ i (t) Tx Packetizer Quality-Aware Compressor c 1,1 c 1,2 c i,1 c i,1 λ i (t) …… …… c 1,|C 1 | c i,|C i | Video v 1 Video v i Chunk Fetching Video Storage Transmitter Receiver Queue Backlog Size, Q i (t) D2D Link, l i (t) Streaming Arrival Process: Placement of Chunks Streaming Departure Process: Transmission of Bits Scheduling: Message-Passing for MWIS Tx Rx Joint Scheduling and Streaming for Device-to-Device Video Delivery A Device Model A Link Model Time Unit Time (t) for (i) Scheduling Decision, (ii) Transmission of Bits Unit Time for Chunk Placement (τ) …… MWIS-Based Scheduling Formulation Computing the Solutions of MWIS-Based Scheduling via Message-Passing Streaming (Arrival Process: Placement of Chunks) Streaming (Departure Process: TX of Bits) Objective Function: maximization of the total quality over all links subject to the rate stability of all scheduled transmitter queues Stochastic Optimization Each Chunk Unit Time, choose q, PSNR of Current Chunk with q i (t) Bitrates of Current Chunk with q i (t) Device-to-Device (D2D) Actively discussing in next-generation cellular standards (3GPP) FlashLinQ (Allerton 2010), one of several proposals for LTE-Direct FlashLinQ The most well-known scheduling algorithm in D2D systems However, (1) It does not contain video-aware operation, (2) It does not consider queue dynamics in D2D transmitters. Scheduling with Max-Weight Independent Set (MWIS) Formulation Streaming with Quality-Awareness subject to Queue Rate Stability Joint Scheduling and Streaming Sanghavi and Shah NIPS’07

Upload: others

Post on 28-Dec-2021

5 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Joint Scheduling and Stochastic Streaming for Device-to

Joint Scheduling and Stochastic Streaming for Device-to-Device Video Delivery

Joongheon Kim (USC), Andrea Turci (U. Bologna), Giuseppe Caire (USC), Andreas F. Molisch (USC) E-mails: [email protected], [email protected], [email protected], [email protected]

Introduction

IEEE Information Theory and Applications Workshop (ITA) Graduation Day Talk (February 12, 2014)

University of California, San Diego (UCSD), La Jolla, California, USA

Streaming with Stochastic Network Optimization Scheduling with MWIS Formulation

Transmitter

Qi(t) μi(t)

Tx

Packetizer Quality-Aware

Compressor

c1,1 c1,2

ci,1 ci,1

λi(t)

……

……

c1,|C1|

ci,|Ci|

Video v1

Video vi

… Chunk

Fetching Video

Storage

Transmitter Receiver

Queue Backlog Size, Qi(t)

D2D Link, li(t)

Streaming – Arrival Process:

Placement of Chunks

Streaming – Departure Process:

Transmission of Bits

Scheduling:

Message-Passing for MWIS

Tx Rx

Joint Scheduling and Streaming for Device-to-Device Video Delivery

A Device Model A Link Model

Time

Unit Time (t) for

(i) Scheduling Decision, (ii) Transmission of Bits

Unit Time for

Chunk Placement (τ)

……

MWIS-Based

Scheduling Formulation

Computing the

Solutions of

MWIS-Based Scheduling

via Message-Passing

Streaming (Arrival

Process: Placement of

Chunks)

Streaming (Departure

Process: TX of Bits)

Objective Function:

maximization of the total quality over all links subject to

the rate stability of all scheduled transmitter queues

Stochastic

Optimization

Each Chunk Unit Time, choose q,

PSNR of Current

Chunk with qi(t)

Bitrates of Current

Chunk with qi(t)

• Device-to-Device (D2D)

Actively discussing in next-generation cellular standards (3GPP)

FlashLinQ (Allerton 2010), one of several proposals for LTE-Direct

• FlashLinQ

The most well-known scheduling algorithm in D2D systems

However,

(1) It does not contain video-aware operation,

(2) It does not consider queue dynamics in D2D transmitters.

Scheduling with

Max-Weight Independent Set

(MWIS) Formulation

Streaming with

Quality-Awareness subject to

Queue Rate Stability

Joint Scheduling and Streaming

Sanghavi and Shah

NIPS’07