The Think Layer Additional Material - Lecture Slides | ECE 284, Study notes of Electrical and Electronics Engineering

Material Type: Notes; Class: Topics/Computer Engineering; Subject: Electrical & Computer Engineer; University: University of California - San Diego; Term: Fall 2005;

Typology: Study notes

Pre 2010

Uploaded on 03/28/2010

koofers-user-jhb-1
koofers-user-jhb-1 🇺🇸

1

(1)

10 documents

1 / 3

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
1
1
ECE 284
ECE 284
The Link Layer
Additional Material
This material will not be part of the exam.
It is included here for completeness.
2
ECE 284
ECE 284
Leaky Bucket Mechanism
Leaky Bucket Mechanism
Leaky bucket: traffic policing
Bucket holds up to
B
tokens
b
is the number of tokens currently in the bucket
Tokens fill up the bucket at rate
λ
Transferring a packet from the input queue consumes a token from the bucket
The number of packets admitted in a time int erval is bounded:
Variants
Packet-level leaky bucket
Bit-level leaky bucket
λ
Bb
(
)
BttttD + 122,1 ][
λ
3
ECE 284
ECE 284
Wireless Fair Queuing
Wireless Fair Queuing
Goal: provide fairness despite temporary localized channel errors
Channel errors should be almost transparent to the user
If a scheduled flow is experiences a bad channel, one with a good channel will
be allowed to transmit instead to avoid wasting capacity
Flows most be compensated for capacity lost due to channel errors
Issues:
Monitoring the channel state
Separation between flows: flows that always percei ve a good channel should
not be impacted
4
ECE 284
ECE 284
Building Blocks of Wireless Fair Queuing
Building Blocks of Wireless Fair Queuing
Error free service model: used as a reference
Lead: ahead of error free service, received more band width so far
Number of slots that have to be relinquished in the future
Lag:
Definition 1: difference between error-free and actual service (independent
whether another flow could transmit or not)
Definition 2: number of slots during which another flow transmitted instead
Compensation model: when are slots of a leading flow relinquished and how
are they distributed amongst lagging flows
Giving up lead: all at once or more graceful degradation
Distributing lag: preferential or fairly (proportional to its lag)
Decouple slot queues from packet qu eues
Slot are scheduled, based on packet arrivals
A separate mechanism selects which packet is transmitted
5
ECE 284
ECE 284
Wireless Fair Service (WFS)
Wireless Fair Service (WFS)
A variant of WFQ is used as the reference model
Separate rate weight
wi
and delay weight
φi
Decouple rate and delay constraints
Lead:
l
i
number of extra slots received
Lag:
b
i
number of slots used by other flows
In sync: allocation same as error free service, neit her leading nor lagging
+=
i
i
ii w
kL
kSRkS )1(
)1(,max)(
i
i
ii
kL
kSkF
φ
)(
)()( +=
6
ECE 284
ECE 284
WFS Behavior
WFS Behavior
Compensation model:
Leading flow relinquishes of its slots
Lagging flow receives a fraction of the relinquished slots
Properties
Short-term and long-term fair
Achieves delay and throughput bounds
Delay and bandwidth decoupling
Graceful relinquishing of slots
Fair compensation of lagging flows
max
i
il
l
kk
ib
b
pf3

Partial preview of the text

Download The Think Layer Additional Material - Lecture Slides | ECE 284 and more Study notes Electrical and Electronics Engineering in PDF only on Docsity!

ECE 284ECE 284^1

The Link Layer

Additional Material

This material will not be part of the exam.

It is included here for completeness.

ECE 284ECE 284^2

Leaky Bucket MechanismLeaky Bucket Mechanism

„ Leaky bucket: traffic policing

● Bucket holds up toB tokens

● b is the number of tokens currently in the bucket

● Tokens fill up the bucket at rate λ

● Transferring a packet from the input queue consumes a token from the bucket

„ The number of packets admitted in a time interval is bounded:

„ Variants

● Packet-level leaky bucket

● Bit-level leaky bucket

B

b

D [ t 1 , t 2 ]≤ λ⋅ ( t 2 − t 1 ) + B

ECE 284ECE 284^3

Wireless Fair Queuing Wireless Fair Queuing

„ Goal: provide fairness despite temporary localized channel errors

● Channel errors should be almost transparent to the user

„ If a scheduled flow is experiences a bad channel, one with a good channel will

be allowed to transmit instead to avoid wasting capacity

„ Flows most be compensated for capacity lost due to channel errors

„ Issues:

● Monitoring the channel state

● Separation between flows: flows that always perceive a good channel should

not be impacted

ECE 284ECE 284^4

Building Blocks of Wireless Fair QueuingBuilding Blocks of Wireless Fair Queuing

„ Error free service model: used as a reference

„ Lead: ahead of error free service, received more bandwidth so far

● Number of slots that have to be relinquished in the future

„ Lag:

● Definition 1: difference between error-free and actual service (independent

whether another flow could transmit or not)

● Definition 2: number of slots during which another flow transmitted instead

„ Compensation model: when are slots of a leading flow relinquished and how

are they distributed amongst lagging flows

● Giving up lead: all at once or more graceful degradation

● Distributing lag: preferential or fairly (proportional to its lag)

„ Decouple slot queues from packet queues

● Slot are scheduled, based on packet arrivals

● A separate mechanism selects which packet is transmitted

ECE 284ECE 284^5

Wireless Fair Service (WFS) Wireless Fair Service (WFS)

„ A variant of WFQ is used as the reference model

● Separate rate weightw i and delay weight φi

● Decouple rate and delay constraints

„ Lead:li number of extra slots received

„ Lag:b i number of slots used by other flows

„ In sync: allocation same as error free service, neither leading nor lagging

i

i i i

w

L k

S k RSk

() max , ( 1 )

i

Fik Sik Lik

ECE 284ECE 284^6

WFS Behavior WFS Behavior

„ Compensation model:

● Leading flow relinquishes of its slots

● Lagging flow receives a fraction of the relinquished slots

„ Properties

● Short-term and long-term fair

● Achieves delay and throughput bounds

● Delay and bandwidth decoupling

● Graceful relinquishing of slots

● Fair compensation of lagging flows

max i

i

l

l

k k

i

b

b

ECE 284ECE 284^7

Need for Adaptation Need for Adaptation

„ Intrinsic variability

● Channel variability

Š Dynamic channel conditions in space

and time

● Traffic variability

Š New streams, applications, users

Š Non-constant data rate, e.g. mpeg

video

„ Traditional approach: design system for

worst case operating conditions

● E.g. certain outage probability

● Wasteful in terms of resources

„ Adaptation

● Adjust the system settings to the best

operating point for the current

channel/traffic conditions and constraints

Static snapshot of signal intensity map

Source:[IEEE99] (^) ECE 284ECE 284^8

Packet Size AdaptationPacket Size Adaptation

„ Variability

● Channel fading rate varies: two-state Markov chain for different speeds (~transition

prob) and same BERgood

„ Control knobs

● Adjust the packet length

● Fixed transmit power and raw data rate

● TCP traffic source

● Erroneous packets are retransmitted at the MAC level

● Include error coding

„ Minimize the energy

Sampling effect: fit more packets in the ‘good’ state

Adaptation problem 3

ECE 284ECE 284^9

Packet Size Adaptation Packet Size Adaptation

„ Two observations

● For a particular speed, there is an optimal packet size (depends only on BERgood)

● At large packet sizes, the sampling effect causes a dependency on mobile speed

L (bytes)

Normalized energy

(as adaptation problem 2)

ECE 284ECE 284^10

Adaptive Error Coding [Let99] Adaptive Error Coding [Let99]

„ Variability

● Channel quality varies: two-state Markov chain with same speed, varying BERgood

„ Control knobs

● Adjust the code rate of a RS coder to compensate for BER variations

● Fixed transmit power and raw data rate

● TCP traffic source; erroneous packets are retransmitted at the MAC level

„ Minimize the energy per useful bit

BERgood

Energy per good bit ( μ J/bit)

Adaptation problem 4

ECE 284ECE 284^11

„ ILP problem formulation in case of perfect knowledge

● SendR tot bits within delay boundT A (symbol rate is fixed)

● Minimize the total energy consumption, by selecting the rateR i for each symboli fromJ

possible settings {R 1 , ..R J}

Analysis of the RadioAnalysis of the Radio

E x e i

J j

i = ∑ ij ⋅ ij ∀

= 1

R x R i

J j

i = ∑ ij ⋅ j ∀

= 1

S i

i

= ∑= Rtot 1 R

S i

i

min Etot 1 E

x i

J j

∑ ij = ∀

=

0

x ij ∈{ 0 , 1 } ∀ i , j

I( 0 )

0

⎥⋅^ ≠

= =⎡^ ⋅ ⋅ +

E j b j j i

S S

ij N C R

ER

C R

R

e P

S = TA ⋅ R S

1 2 S

T A

ECE 284ECE 284^12

Scaling and SchedulingScaling and Scheduling

Queue Sender Receiver

Queuing Practical

Start of transmission

„ Energy-aware queuing system: DMS with constant channel [Sch01]

● Principle: transmit slower when there are fewer packets in the buffer

Scaling problem 6-

T av ( μ s)

Eav ( μ J)

bn

S n