On LinkLayer Reliability and Stability for Wireless Communication

1
On Link-Layer Reliability and Stability for
Wireless Communication

• Paper by Sohraab Soltani, Kiran Misra, and
  Hayder Radha
• Presenter Ke Gao
• Instructor Professor Beyah

2
Overview

• Introduction
• Channel Model and Distortion Model
• ACE Code Embedding Rate
• ACE Protocol
• Performance Evaluation

3
Motivation

• Popular wireless link-layer protocols achieve
  some level of reliability using Automatic
  Repeat Request (ARQ ) or Hybrid ARQ mechanisms.
• However, they ignore the stability aspect of
  wireless communication, and rely on higher layers
  to provide stable traffic flow control.
• The paper proposes Automatic Code Embedding (ACE)
  of link layer protocol which both reliability and
  stability are targeted in wireless networks.

4
Related Work

• ARQ-based Schemes (IEEE 802.11)
• Incorporates frame check sequence (FCS) to detect
  errors and automatic repeat request (ARQ) to
  retransmit corrupted packets.
• Network utilization deteriorates steadily and
  rapidly with increasing channel Bit Error Rate
  (BER).

5
Related Work - cont

• HARQ-based Schemes
• The sender encodes the packet payload with an
  error-correction code prior to the transmission.
• The receiver requests for a retransmission when
  the decoding of the received packet fails.
• However unlike ACE, the HARQ is not adaptive with
  respect to channel condition and does not address
  throughput stability issues raised by varying
  traffic demand.

6
Related Work - cont

• Cross-Layer Approach
• In recent years, many papers in multimedia
  applications have proposed cross layer mechanisms
  to overcome performance limitations imposed by
  conventional protocols.
• A significant drawback of the cross-layer
  protocols is that their implementations require
  major modifications in transport and application
  layers.

7
Channel Model

• A channel model describes the process under which
  errors are introduced in a transmitted packet
  over a wireless link.
• During the ith transmission interval, a message
  is transmitted over a Binary Symmetric Channel
  (BSC) with cross-over BER ei.
• We can model a wireless channel as a discrete
  Markov chain with N states where each state is a
  representation of a BSC with a particular BER

8

• Assume a homogenous and stationary Markovchain
  with transitional probability matrix P and the
  limiting probabilities p (p1, , pN ).

9

• The channel capacity gives an upper bound on the
  average (reliable) information transmission rate
  for the wireless channel under consideration.

10
Distortion Model

• The distortion model measures the distortion
  level of a received packet and computes the
  likelihood of successful recovery of the packet
  under embedded channel coding.
• This density determines the likelihood of
  successfully error recovery using a-error
  correcting codes.

11
ACE Code Embedding Rate

• Code Rate Reliability (upper bond)
• Definition 1. System is reliable when
  information is transmitted with no or diminishing
  error over a wireless channel.
• Lemma 1. The operational code embedding rate that
  ensures reliability in wireless transmission over
  a channel in state Si is bounded above by

12
ACE Code Embedding Rate - cont

• Code Rate Stability
• Definition 2. System is stable when higher layers
  are neither starved for information packets nor
  is there a glut of packets leading to buffer
  overflow.
• Lemma 2. The operational code embedding rate that
  ensures stability in wireless transmission over a
  channel in state Si has a lower bound

13
(No Transcript)
14
(No Transcript)
15

• Optimal solution for code embedding rate that
  ensures reliability and stability conditions

16
ACP Protocol
17
ACP Protocol - cont

• ACE utilizes the receiver channel estimate and
  buffer fags to assess the status of the channel
  condition and the receiver buffer. Depending on
  this assessment, ACE determines the composition
  of the next message to be transmitted by the
  sender.
• In every transmission interval, the receiver
  adjusts the parameters of the channel model and
  sends its estimate of the current channel
  condition in an acknowledgment message.

18
Performance Evaluation
19
(No Transcript)
20
Real-time Traffic
21
Non-Realtime Traf?c

• Figure 8 The average goodput of ACE and
  IEEE802.11 ARQ over various channel conditions.
  Note that channel capacity in each FIgure
  represents the maximum amount of achievable
  goodput without errors.

22
Throughput analysis of TCP
23
Real-Time Video Simulation
24
Thank You.

• Q A