resources and …cs778/ralf/08b-qos.pdf · kom - industrial process and system...
Post on 09-Sep-2018
220 Views
Preview:
TRANSCRIPT
u-darmstadt.de Fax. +49 6151 16-3052
, Dept. of Computer Science51 166151,M.tu-darmstadt.de Fax. +49 6151 166152
ogy
08B-QoS.fm 1 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser Multimedia-Systems:
Resources andQuality of Service (QoS)
Prof. Dr.-Ing. Ralf SteinmetzProf. Dr. Max MühlhäuserMM: TU Darmstadt - Darmstadt University of Technology,
Dept. of of Computer ScienceTK - Telecooperation, Tel.+49 6151 16-3709,Alexanderstr. 6, D-64283 Darmstadt, Germany, max@informatik.t
RS: TU Darmstadt - Darmstadt University of Technology,Dept. of Electrical Engineering and Information Technology
KOM - Industrial Process and System Communications, Tel.+49 61Merckstr. 25, D-64283 Darmstadt, Germany, Ralf.Steinmetz@KOGMD -German National Research Center for Information Technolhttc - Hessian Telemedia Technology Competence-Center e.V
User Interfaces
.Synchro-nization
GroupCommuni-
cations
Programming
ms Communications
ice Networks
ession
Video Audio
08B-QoS.fm 2 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
Scope
Usa
ge Applications
Learning & Teaching Design
Ser
vice
s ContentProcess-
ing
Docu-ments
Security ..S
yste
ms Databases
Media-Server Operating Syste
Opt. Memories Quality of Serv
Bas
ics Computer
Archi-tectures
Compr
Image &Graphics
Animation
stems
se
08B-QoS.fm 3 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
Contents
1. Motivation
2. Characteristics of Real-time / Multimedia Sy
3. QoS - Definition
4. Resources
5. Providing QoS
Resource Management Phases
5.1 QoS Provisioning - Setup Phase
5.2 QoS Provisioning - Data Processing Pha
6. QoS Architectures
7. Conclusion
local:•harddisk
recording•interactive DVD•computer basedtraining
distributed•conferencing•video on demand•IP-Telephony
how to describe that?
emory
ing
tion
m
re
08B-QoS.fm 4 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
1. Motivation
basically, we deal with two types of systems:
Basic terminology• Resources• Realtime• Quality of Service
What and how much of it do we need and
Network
Network
MCPU
Receiv
Applica
Syste
SoftwaMemoryCPU
Sending
Application
System
Software
MemoryCPU
System
Software
em:
:
08B-QoS.fm 5 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
Motivation (cont.)
When we know about the needs, how to fulfill th• A QoS model and its implications
• QoS specification• QoS calculation• QoS enforcement
QoS has different implications in different fields• Operating system / Resource scheduling• File system organization• Compression• Communication system support• Media synchronization• ...• User Interface
timedia Systems
mputation depends notupon providing the
e processing in a given
ictable intervals
ion requests from the OS
ically or spontaneousnstraints !
08B-QoS.fm 6 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
2. Characteristics of Real-time / Mul
Real-time System:
“A system in which the correctness of a coonly on obtaining the right result, but alsoresult on time.”
Real-time Process:
“A process which delivers the results of thtime-span.”
Real-time Application - examples• Control of temperature in a chemical plant
• driven by interrupts from external devices• these interrupts occur at irregular and unpred
• Example: Control of a flight simulator• execution at periodic intervals• scheduled by timer-service which the applicat
Common characteristics:• internal and external events that occur period• correctness also depends on meeting time co
to finish an operation, e.g.
value for the useric) system failure
ut not too many deadlines are
ser
Time
ess
08B-QoS.fm 7 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
Deadlines in Realtime Systems
A deadline represents the latest acceptable timefor the presentation of a processing result
• Hard deadlines:• should never be violated• result presented too late after deadline has no• violation means severe (potentially catastroph• Example: Nuclear power plant
• Soft deadlines:• deadlines are not missed by much• in some cases the deadline may be missed, b
missed• presented result has still some value for the u• Example: train/plain arrival-departure
Start process Deadline to finish proc
hard
soft?
ation
nts
ynchronisation
08B-QoS.fm 8 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
Realtime System - Requirements
Primary goal:• deterministic behaviour according to specific
(results in a variety of requirements)
mandatory requirements:• Predictable (fast) handling of time-critical eve• Adequate schedulability• Stability under overload conditions
desirable requirements:• Multi-tasking capabilities• Short interrupt latency• Fast context switching• Control of memory management• Proper scheduling• Fine-granularity of timer services• Rich set of interprocess communication and s
mechanisms
special characteristics:
proper handlingssion / Video)
s are often not evident
eadline not met, ...)it? just "know" ...?)
ction, VRML etc.)
08B-QoS.fm 9 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
Multimedia Systems
New application area for Realtime systems with• Typically soft real-time and not (that) critical• Requirements may often be adapted to ensure
• e.g. Scalability (remember lecture on Compre
Note: when reading literature, some distinction• level of enforcement
(guarantee, best-effort, "proper" handling if d• "exception" handling (skip on to next data un
Characteristics:• Periodic processing
(increasing importance of non-periodic: intera• Large bandwidth• End-to-End Guarantees• Fault-tolerance• Fairness• Standardization
ice
("Measure"?)
ional audio):
)
r, it does so as a "service"
08B-QoS.fm 10 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
Quality of Media vs. Quality of Serv
Quality.... with respect to a Quality "Parameter"
Example Audio QoM (huge set exists in profess• Frequency Spectrum (linear amplification? ...)• Signal2Noise Ratio SNR (noise, click, ...)
Example Video QoM• spatial / temporal resolution• SNR
intuitively spoken:• QoM: something like "HiFi Audio" (16..20k Hz• QoS: something like "Hi bandwidth" (1Gbps)
but: if application delivers HiFi Audio to the use
therefore in the remainder: QoS
system according to
Transport System..
...
08B-QoS.fm 11 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
3. QoS - Definition
Quality of Service =
„well-defined and controllable behavior of aquantitatively measurable parameters“
Layer model:
Different Service Objects:• Media, Media Streams --> Packet Streams• Tasks• Memory areas
User
Application
MM System
File System Local Processing ....
......
ption
08B-QoS.fm 12 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
QoS - Layer Model
Examples: both qualitative / quantitative descri
Perception QoS• Tolerable Synchronisation Drift• Visual Perceptability
Application QoS• Media Parameters• Media (Transmission) Characteristics
System QoS• CPU Rate / Usage• Available Memory
Communication QoS• Packet Size / Rate• Bandwidth• End-to-End-Delay
Device QoS• Seek / Data Transfer Rate• Sample Rate / Resolution
rt System
System are:
ss / Reliability
08B-QoS.fm 13 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
QoS Parameters - Example Transpo
Common parameters concerning the Transport• Throughput• Delay / Jitter• Loss / Reliability
but also:• Security• Costs• Stability (Resilience)
Lo
Delay
Throughput
f one packetion (-->arrival!) timesl time,
per time interval
sesme intervally lost packets
08B-QoS.fm 14 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
Example QoS Parameters (cont.)
Delay:• Maximum end-to-end delay for transmission o• Delay jitter = maximum variance of transmiss
(math. definition: relative to "expected" arrivaNOT to arrival time of other packets)
Throughput:• Maximum long-term rate
= maximum amount of data units transmitted(e.g. packets or bytes per second)
• Maximum burst size• Maximum packet size
Loss:• Sensitivity class: ignore / indicate / correct los• Loss rate = maximum number of losses per ti• Loss size = maximum number of consecutive
l process
rst-length, ....
ropability)
n ...”
08B-QoS.fm 15 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
Service Classes
note: QoS parameters often subject to statistica
---> mean, min, max, distribution, variance, bu
Guaranteed Service• values or intervals of QoS parameters
• deterministic (at any time)• statistical (consider a time interval or certain p
QoSmin <= P <= QoSmax
Predictable Service• consider history
• from the very beginning of calculation• in a shifting time window
• “if it was like that in the last ..., you can rely o
Best Effort Service• no or just partial guarantees
icerce consumption / costs
increasingQoS qualityoo good
08B-QoS.fm 16 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
QoS Intervals
Parameter values result in• acceptable regions• inacceptable regions
of QoS
(here:) in one-dimensional intervals
note:• below required QoS level - no reasonable serv• above required QoS level - unnecessary resou
desired QoSrequired QoS
application’s needs:
QoS too bad QoS tacceptable QoS
min. QoS-Frame-Rate.
max. QoS-Frame-Rate.
oStion
640x480
Resolution (Pixel)
08B-QoS.fm 17 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
QoS intervals
also: multidimensional intervals
Unacceptable Region
min. QoS-
Acceptable
Region
max. QResoluResolution
Frame Rate
10
30
20
0
80x40
08B-QoS.fm 18 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
4. Resources
Classification
by functionality• active resources
• actively fulfill a certain task• e.g. processor, network adapter
• passive resources• provide “space”• e.g. memory, frequency spectrum, filesystem
by availability for concurrent usage• exclusive• shared
by occurence• single• multiple
Common parameter:• Capacity - allows quantitative description
hardware
in year X
danturces
2000
resources
utources
indow of Scarcity”derson et al., 1990]
08B-QoS.fm 19 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
Resources - Availability
Starting point:• scarce, but sufficient resources
Goal:• provide best service at lowest possible costs
Conclussion• need for resource management
requirements
abunreso
insufficientresources
1980 1990
interactivevideo
high-qualityaudio
networkfile access
sufficient bscarce res
“W[An
QoSafter
processing
Data out
08B-QoS.fm 20 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
Relationship: QoS - Resources
Model:
Resource with
certain capacity
Processing, using a
QoSbefore
processing
Data in
Resource SpecificInformation
Reservation
Database
08B-QoS.fm 21 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
Architecture
Resource
Scheduler
Resource
Manager
CONTROL
DATA
Resource
Monitor
resourcereservation
QoS guaranteesto user
y resource schedulinging, adaptation
08B-QoS.fm 22 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
5. Providing QoS
Resource Management Phases
user’s QoSrequirements
admission control
phase 1 (Setup):
data arrivalon streams
QoS enforcement bshaping, loss handl
phase 2 (Data processing):
rejection
negotiation
calculation
of QoSguarantees
tations
resourcereservation
QoS guaranteesto user
08B-QoS.fm 23 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
5.1 QoS Provisioning - Setup Phase
Definition of required parameters• implicit or explicit by application or user
Distribution and Negotiation
Translation between different layers• especially if they use different semantics / no
Transformation• QoS parameter => Resource requirements
Allocation and coordination of resources• along path(s) source(s) => sink(s)
user’s QoSrequirements
admission control
phase 1 (Setup):
rejection
negotiation
calculation
of QoSguarantees
User (Callee)
System (Callee)
plication (Callee)
p
08B-QoS.fm 24 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
QoS Calculation and Negotiation
Model:
User (Caller)
System (Caller)
Application (Caller) Ap
Service User
Peer-to-Peer-Relationshi
(Caller-to-Callee)
Service Provider
oS parameters(lower) value(s) on confirm
e
QoS-ValueChanged
ResponseConnect
QoSaveconfirm
eq
t3
08B-QoS.fm 25 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
QoS Negotiation
bilateral peer-to-peer• service provider may not modify requested Q• only service user at receiver side may modify
Service Provider
Caller
Peer-to-Peer
Negotiation
Calle
QoSavereq
QoSaver
ResponseConnect t4 Response
Connect t1 ResponseConnect t2
QoS-ValueRequested
can’t meet it
r-to-layer-negotiation=> heterogenity of receivers
limited target value
ered distrib. services:reduces QoS parametersoS, initator accepts/rejects" battle, see networking)
08B-QoS.fm 26 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
QoS Negotiation (cont.)
bilateral layer-to-layer• only between adjacent parts
• between local service users and providers• between sender and network
unilateral• no modification of requested QoS parameters• but just accept or reject• receiver may accept QoS parameter though it
• example: color TV broadcast
hybrid• uses unilateral mode at a certain bilateral laye
• example: broadcast/multicast communication
further:• trilateral for information exchange or: ... for a
commonly accepted for connection-oriented lay• initiator requests QoS, each entity accepts OR• final "Connect-Confirm" w/ overall minimum Q• (NOTE: "connection-oriented" vs. "flow-based
ll be available
use and new request(s) is
back” to prevent users
08B-QoS.fm 27 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
Admission Control
check, whether requested resources are and wi
especially important for shared resources:• CPU• network paths• buffer space: memory
simple rule:
check whether sum of resources already inless or equal available resource capacity
... may be applied to• availability of buffers (spatial)• bandwidth (net)
more complex: check for schedulability
note:• strong relationship with Pricing / Billing• efficient mechanisms will use “economic feed
from requesting whatever they can get
f QoS guarantees
nditions (e.g. CPU load)
reservedforappl. 1
reservedforappl. 2
time
reservedforappl. 1reservedforappl. 2
lict
time
08B-QoS.fm 28 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
Resource Reservation
Fundamental concept for reliable enforcement o• pessimistic - results in Quaranteed QoS
• optimistic - results in statistical QoS
• may use monitoring and react on overload co
unused
guaranteedQoS:
unused
needs of appl. 1
needs of appl. 2
needs of appl. 1
needs of appl. 2
statisticalQoS:
conf
ample
of aspects
08B-QoS.fm 29 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
Resource Reservation Aspects - Ex
Example: Communication System ===> variety
Reservation Model• Sender-initiated• Receiver-initiated• Explicit vs. Implicit• Out-of-Band vs. In-Band
Reservation Style• Semantics and Notation• Heterogenity and Multicast-Support
Reservation Protocols• IP V.5: ST-II• RSVP (Resource reSerVation Protocol)
sing Phase
istics of processed data)
y resource schedulingling, adaptation
08B-QoS.fm 30 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
5.2 QoS Provisioning - Data Proces
maintain resource reservations
use:• adequate traffic shaping (to ensure character• scheduling• feedback and adaption
mechanisms
note: concurrent requests !
data arrivalon streams
QoS enforcement bshaping, loss hand
phase 2 (Data processing):
ugh guarantees could be
t
after Shaping
08B-QoS.fm 31 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
Shaping
Characteristics of Multimedia Traffic• bursty (remember lecture Compression)• concurrent requests may cause problems tho
met (e.g. buffer overflow)
Basic principle
r(t)
average Cell Rate
max Cell Rate
Original Source
(r,T) shaping
rop) and possible delay
gned per connectionan r Bits
next interval (not economic)
ed (supply)
Data Capacity
onstant
08B-QoS.fm 32 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
Shaping - Leaky Bucket Algorithm,
Leaky Bucket: Bucket Size• determines maximum capacity till overflow (d
(r, T) shaping:• frames of T bits (system-wide), fraction r assi• within interval T, sender may not send more th• if "current packet" would exceed r -> wait for
Data to be transmitt
Regulation RN
c
supply
LeakyBucket
constant
pacity βs (<=β) usedmore than β+R*t tokens)
e Controled into network
cket
acity b
ts
08B-QoS.fm 33 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
Shaping: Other Algorithms
Token Bucket Algorithm• tokens drop into bucket with rate R, bucket ca• packet (="burst") size determines no. of token• effect: limited-size bursts allowed (within t, no
Token Bucket Algorithm with Leaky Bucket Rat• problem w/ token bucket: burst directly releas• therefore, put "burst" into another (leaky!) bu• --> burst drop out of leaky bucket smoothly
tokens (rate R)
data to be transmitted
token cap
limited burs
LeakyBucket
am stalls during retransmit)RC) for detection of error onlyacknowledged packets resent
erroneous Xmit)r packets until reXmit)sonable" no. of packets)solution" for multimedia)undancy for correction!f FEC and CRC
esentation!i-speed LANsice-....: re-buffer after pause
08B-QoS.fm 34 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
Loss Handling
Error Detection• by means of redundancy / checks / analysis
Loss Handling
2 basic categories:• (partial) Retransmission (problem: media stre
• in-packet redundancy (cyclic redund. check C• automatic repeat request ARQ: non (or neg.)• three facets:• Go-back-N (error -> "reset" link to state before• Selective Retransmission (error: buffer furthe• Using partially error-free streams (reXmit "rea
• Prevention (formerly considered "only viable• Forward Error Correction (FEC): in-packet red• Priority Coding: --> mix of FEC "degrees" or o
• Multimedia-proof retransmission:• Slack ARQ: buffer received packets before pr• buffer time long enough to support reXmit in h• for, e.g., voice streams of type voice-pause-vo
esources
ctions to reduce generated
nder to slow down
o sender
Monitor/Decode
08B-QoS.fm 35 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
Adaption - Feedback Control
Monitor load of network and local end-system r
If significant changes occur, take appropriate aload
• Explicit communication – receiver informs se• Completely in network on a hop-by-hop basis• By feedback from congested network nodes t
Variety of possible reactions• e.g. Layered transmission• Degredation, ...
Code/ScaleLess/More
Network
ed handling of Flows in
ort heterogenous needs
vel Agreements and ensuredescribed
/ dedicated treatment of flows
08B-QoS.fm 36 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
6. QoS Architectures
Examples (Communication Layer)
• Heidelberg Transport System (HeiTS)• uses ST-II (IPv5)
• Internet Integrated Services• use existing infrastructure, but deploy dedicat
Transfer System• Resource Reservation Protocol RSVP to supp
• Differentiated Service• Granularity based on TOS IP Header Field• Define Service Classes, Negotiate Service Le
dedicated treatment of Flows that behave as
• IPv6• QoS support as one design criteria• dedicated header fields to allow classification
ovisioning in 2 phases:
aling ...)
08B-QoS.fm 37 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
7. Conclusion
Realtime- and Multimedia Systems
Quality of Service - Definition and Concepts
Resources and Resource Management / QoS pr• QoS specification, calculation and negotiation
• Requirements of the application• Functions to calculate QoS guarantees• Guarantees returned by the system• Reservation of resource capacities
• QoS enforcement:• Scheduling of resource access• Monitoring and adequate actions (shaping, sc
s, netowork
)) packetsicaste in full resolutions periodically
receiver 1
receiver 2
receiver 1
receiver 2
08B-QoS.fm 38 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
Outlook: Communication Protocols
Reservation / QoS initiated by: sender, receiver
example RSVP: receiver-initiated• sender sets up paths (requested by receivers?• receivers periodically send reservation (RSVP• paths (~flows) may, e.g., represent video mult• receivers might not be able to present / receiv• receivers may just drop video -> RSVP packet•
Internet(Multicast)
sender 1
data (session)PATH
RESV
RSVP
RSVP
RSVPInternet
(Multicast)
sender 1
data (session)PATH
RESV
RSVPRSVP
RSVPRSVP
RSVPRSVP
es (cf. chapter MM sync.)
ion
08B-QoS.fm 39 20.June.01
Scope
Contents
htt
p:/
/ww
w.k
om
.e-t
ech
nik
.tu
-da
rmst
adt
.de
htt
p:/
/ww
w.t
k.in
form
atik
.tu
-da
rmst
adt
.de
©R
.S
tein
me
tz,
M.
Mü
hlh
äu
ser
QoS vs. Human Perception
Rule-of-Thumb: Eye integrates, Ear differenciat
--> tolerant visual perception 1: overlay separat
--> tolerant visual perception 2: completion
top related