Transcript Admissible traffic load of real time class of service for inter

Admissible Traffic Load of
Real Time Class of Service for
Inter-domain Peers
Halina Tarasiuk, Robert Janowski and Wojciech Burakowski
Warsaw University of Technology, Poland
Contents
• Classes of service concept as an approach for providing
strict QoS guarantees at the network level
– Experiences from AQUILA project (5FR)
– EuQoS project (6FR – in progress)
• RT service at an inter-domain peer
• CAC for RT service
– Algorithm
– Numerical results
• Summary
Classes of service concept as an approach
for providing strict QoS guarantees at the
network level
– Experiences from AQUILA project (5FR)
– EuQoS project (6FR – in progress)
QoS at different levels
User level
Subjective
assessment
User level
Subjective
assesment
user
user
ITU G.1010
codec
codec
Application
level
Application
level
Additional
mechanisms
Network
level
Additional mechanisms
(e.g. playback buffer)
ITU Y.1541
Network
interface
Network
interface
Network
level
Classes of
Service
To guarantee packet
losses, packet delays
Class of service concept
• A „service class” represents a set of traffic that
requires specific delay, loss and jitter
characteristics from the network for which a
consistent and defined per hop-behaviour applies
• A service class pertains to applications with
similar characteristics and performance
requirements
Discussed Classes of services (IETF proposal)
InterProvider
Service Class
(Aggregate)
Ctrl
Real Time
Tolerance To
Loss
Delay
Jitter
Low
Low
Yes
VLow VLow VLow
Tolerance To
End-To-End
PHB
Service Class
CS
EF
None Real
Time
Low
LIM
Yes
AF
Best Effort
NS
NS
NS
DF
Loss
Delay
Jitter
DSCP
Name
DSCP
Value
Network
Low Low
Yes
CS7 111000
Control
Telephony VLow Vlow VLow EF
101110
Signalling
Low Low
Yes
CS5 101000
MM
L-M Vlow Low AF4x 100xx0*
Conferencing
RT
Low Vlow Low
CS4 100000
Interactive
Broadcast
VLow Med Low
CS3 011000
Video
MM
L-M Med
Yes AF3x 011xx0*
Streaming
Low Latency
Low L-M
Yes AF2x 010xx0*
Data
OAM
Low Med
Yes
CS2 010000
High ThruPut
Low M-H Yes AF1x 001xx0*
Data
Standard
NS
NS
NS
DF
000000
• End-to-end – related to applications (visible by users)
• Aggregated in some network parts (maintained by the network)
Definition of a service class
1. QoS objectives: values of packet losses, delays...
2. Types of connections: p2p
3 Traffic descriptors: single-, double token bucket,
more advanced
A. Provisioning of resources: static, dynamic
B. CAC: based on declarations, based on
measurements
C. Tuning mechanisms at the packet level (PHB:
classifiers, scheduling, marking, active quieueing..)
Experiences in implementing CoSs -
AQUILA network (2000-2003)
AQUILA(IST-1999-10077)
Adaptive Resource Control for QoS
Using an IP-based Layered Architecture
AQUILA Architecture
Resource Control Layer
Resource Control
Consideration of Network Load
Monitoring
Probing
Results
Resource
Control
Agent
Admission Control
QoS
Request
Core
Router
Edge
Router
Access
Network
Admission
Control Agent
Setting
QoS
Request
QoS Request
Setting
Admission
Control Agent
End-user
Application
Toolkit
Core
Router
Edge
Router
Core
Router
Access
Network
QoS in core networks – IP
prototype solutions: AQUILA
Network service
Traffic
type
Characteristic
examples
Application
example
 Premium CBR
Constant
Small packets
SIP VoIP
low loss, low dela
 Premium VBR
Variable
Large packets
low loss, low
delay
SIP Video
 Premium MM
Adaptive
 Premium MC
Very short
bursts
required
throughput
very low
delay & loss
File transfer
(FTP)
online games
 Standard
Best effort
classical
the rest
Goal: only a few network services
to allow clear service differentiation
Tested CAC algorithm for PCBR service RT service
New flow is admitted if:
(1)
N1
PBRnew   PBRi  C1
i 1
Where N1 denotes the number of connections in progress and
parameter  (<1) specifies the admissible load of capacity
allocated to the PCBR. The value of  can be calculated from the
analysis of M/D/1/B system taking into account the target packet
loss ratio and the buffer size [2].

2 Buffer
2 Buffer  ln( Ploss )
(2)
Where Buffer denotes buffer size in packets and Ploss target packet
loss ratio.
Overall Topology for Trial in AQUILA
Implementing CoSs in EuQoS system
(2004-2006)
End-to-end Quality of Service
support over heterogeneous
networks
Some of the problems to be solved
•
•
•
•
•
Scalable architecture
Signalling system
Providing QoS at the packet level
To cope with network heterogeneity
Etc.
EuQoS Architecture: Physical View
USER 1
Application
EQ-SIP
proxy
EQ-SIP
Signaling
EQ-SDP in
EQ-SIP
proxy
End-to-end QoS EQ-SIP signaling
USER 2
Application
EQ-SIP
Signaling
Virtual Network Layer
Network technology Independent sub-layer
EQ-SDP
EQ-SDP
EQ-NSIS
RM1
EQETP
Protocols
RA1
Access
Network
1
RMi
EQ-NSIS
EQ-NSIS
RMj
EQ-NSIS
RMk
RM2
n
Network technology dependent sub-layer
RAk
RAi
RAj
QoS
Domain
i
QoS
Domain
j
EQ-path
QoS
Domain
k
RAn
Access
Network
2
EQETP
Protocols
EuQoS system
QoS routing – EQ-BGP (path)
QoS signalling – EQ-SSN
QoS Request
QoS Request
QoS Request
R
M
QoS domain path
CoS
RA
QoS Request
R
M
QoS Request
QoS Request
QoS interdomain path
CoS
QoS domain path
QoS interdomain path
CoS
End-to-end CoS path
CoS
R
M
QoS domain path
CoS
Borders
for
Classes
of
service
AC
AC
AC
Ingress BR
AS1
Egress BR
Intra-domain
service in AS1
Ingress BR
Router
input
port
input ports
Ingress BR
AS2
output
port
Inter-domain
service AS1-AS2
output
port
Intra-domain
service in AS2
AC
Intra- and inter-domain Classes of service
AC: admission control
Classes of Service in EuQoS
End-To-End
Classes of
Service
Telephony
RT
Interactive
MM
Streaming
High
ThruPut
Data
Standard
EuQoS Applications (Phase 1)
Medigraf
VTC
VoD
Collabo
data
VTC
ration
transfer
VoIP
X
X
X
X
X
X
QoS Objectives
End-to-end Class of
Service
IPLR
Mean
IPTD
Telephony
10 -3
100 ms
RT Interactive
10 -3
100 ms
MM Streaming
10 -3
1s
U
High Throughput
Data
10 -3
1s
U
U
U
U
Standard
Chat
IPDV
50
ms
50
ms
Type of
connections
P2P
Traffic descriptors
DSCP
code
Peak bit rate, single token
bucket
Peak bit rate, single token
bucket
101110
P2P
Peak bit rate, single token
4
bucket
011xx0 *
P2P
Peak bit rate, single token
4
bucket
001xx0 *
P2P
100000
000000
Plan for developping CoSs in EuQoS
mapping
mapping
User request for
basic end-to-end
CoS
Aggregated CoSs
implemented
inside domain
mapping
mapping
Aggregated
CoSs
implemented
on interdomain link
Aggregated
CoSs
implemented in
core domain
mapping
Aggregated
CoSs
implemented
on interdomain link
Aggregated CoSs
implemented
inside domain
Telephony
RT Interact.
RT
RT
MM Stream.
NRT
NRT
High Thr. D.
BE
BE
Overprovisi
oned
RT
RT
NRT
NRT
BE
BE
Standard
Access 1
(LAN)
Core
Access 2
(UMTS)
Access networks: LAN/Ethernet, xDSL, WiFi, UMTS
IP core: Geant
EuQoS Test Network
Applications vs. Classes of Service
End-to-end
CoSs
EUQOS
Applications
Network
Control
VoIP
Telephony
VTC
RT interactive
VoD
MM
streaming
NRT1
Data
Transfer
High
throughput
data
NRT2
CTRL
CTRL
CTRL
RT
Real Time
RT
Real Time
RT
NRT3
NRT1
CoS on Access CoS in InterNetwork
domain Link
EQ-PATH
EUQOS
Applications
Telephony
VoIP
RT interactive
VTC
MM
streaming
VoD
High
throughput
data
Data
Transfer
NRT1
NRT2
NRT3
End-to-end
CoSs
Network
Control
NRT3
CoS on Acess
Network
• RT service at an inter-domain peer
• CAC for RT service
– Algorithm
– Numerical results
RT Class of Service
• End-to-end Classes of Service
– Telephony for VoIP – short packets (60 bytes)
– RT Ineractive for VTC – long packets (1500 bytes)
• QoS metrics
– IPLR – 10^-3
– Mean IPTD – 100 ms
– IPDV – 50 ms
• Traffic description: single token bucket (PBR, PBRT)
• Policing strategy
– Policing in access network only (entry point); to police (PBR, PBRT)
– We have to define in each access network policy point (node)
Approach 1: not distinguishing between e2e CoSs
• CAC algorithm
N1
PBRnew   PBRi   RT C RT
i 1
 RT 
2 Buff RT
2 Buff RT  ln( IPLR)
• It does not take into account an impact of
packet sizes on IPLR
Approach 2:
Studied system for RT service for
inter-domain pear
Assumptions:
- the input traffic of both end-to-end
CoSs is Poisson process.
- the packet sizes are constant equal
to ‘d1’ and ‘d2’, respectively for
telephony and video conference CoSs
and their ratio (d2/d1) is an integer
denoted by ‘d’.
- the packets of these CoSs enter the
same finite buffer (with buffer size –
Buffer counted in packets).
Analysis (1)
Packet Arrivals
Time
Packet Departures
Q(n+4) Q(n+2)
Q(n+3)
Q(n+1) Q(n-1)
Q(n)
Embedded instants of
inspections of system state
Figure 8. Time evolution of the system state
Where:

Q(n) 1  A  A
if pkt_type1__served
1 2

d
d

Q(n  1)  Q(n) 1   A   A if pkt_type2__served
1
2

i 1
j 1

A A
if system_ was_ empty

1 2
-Q(n) denotes the system state at the end
of n-th embedded time instant
- A1, A2 random variable describing the
number of type 1 (respectively type 2)
packet arrivals during one slot,
-Ratio of packet sizes is denoted as ‘d’
(d2/d1 = d)
Analysis (2)
After some algebra
Q( z ) 
1
( 1d 2  2 d 2 )( z 1) 2
(1d1  2 d1 )( z 1)

e
)

e
(1   ) z
1  2
1  2
1
2
( 1d1  2 d1 )( z 1)
( 1d 2  2 d 2 )( z 1)
e
e
1  2
1  2
 (1   )(e (1d1 2 d1 )( z 1)
z
the load () and the
arrival intensities (1,
2) are related by:
1=1d1=1 (since d1=1);
2=2d2=2d; =1+2
1=w1 ; 2= w2
Eq.9
Analysis (3)
Assuming that the tail probabilities of the queue size distribution
function are well approximated by the dominant pole of Q(z), they
can be written as
1 x
Pr ob{Q  x}  C0 ( )
z0
Further, assuming that the asymptotic constant Co equals 1, the
buffer overflow probability can be expressed as
Ploss  Pr ob{Q  Buffer  1}

1 n
1
1 Buffer1
  C0 ( ) 
( )
z0
z0  1 z0
n  Buffer 2
Eq.12
Analysis (4)
w 
( w1  2 )( z0 1)
we can determine the value of the
w1 
d
z0 
e
w

required decay rate parameter 1/z0.
w1   d2
This decaying rate ensures that the
w2 
( w1d  w2  )( z 0 1)
d
buffer overflow probability will be 
e
0
w2 
w1   d
below target Ploss value.
Eq.14
Steps to calculate the admissible load when all the input parameters (Buffer, Ploss,
d1, d2, percentage contribution of different types of traffic - w1, w2) are given:
1. Given Ploss and Buffer, determine the parameter z0 (Eq.12)
2. Create the equation (14) taking into account the number of traffic types, their
characteristics (intensity, packet sizes) and the assumed input model (Poisson).
3. Solve the equation (14) with respect to  which is the total admissible load.
4. Calculate the admissible load of each traffic class based on the information about
percentage contribution of different traffic classes - w1, w2 (9), i.e. 1=w1, 2=w2
Numerical results (1)
0,9
total admissible load
Ploss
1
1
0.1
0,6
3
5
7
9
11
13
d
15
Target Ploss
I-10% / II-90%
0.01
I-50% / II-50%
I-90% / II-10%
7
0.001
0,3
I-10%/ II-90%
I-50%/ II-50%
0.0001
I-90%/ II-10%
0
d
1
3
5
7
9
11
13
Figure 9. Total admissible load vs.
packet size ratio of two end-to-end
CoSs; target Ploss=10-3
0.00001
15
Figure 10. Packet loss ratio vs.
packet size ratio of two end-toend CoSs; target Ploss=10-3
Numerical results (2)
1
3
5
7
9
11
13
0.1
0.0001
1
1
3
5
7
9
11
13
d
15
0.1
Target Ploss
0.01
0.001
d
15
Ploss
Ploss
1
Target Ploss
0.01
I-10% / II-90%
0.001
I-50% / II-50%
I-90% / II-10%
I-10% / II-90%
I-50% / II-50%
0.0001
I-90% / II-10%
0.00001
Figure 11. Packet loss ratio vs.
packet size ratio of two end-toend CoSs; target Ploss=10-2
0.00001
Figure 12. Packet loss ratio vs.
packet size ratio of two end-to-end
CoSs; target Ploss=10-4
Summary
• QoS guarantees at the network layer we can assure by providing
classes of service
• RT service for inter-domain peers requires adequate CAC algorithm
• The proposed algorithm works correctly and takes into account
differences in packet sizes
• The algorithm will be implemented in EuQoS system and tested
• „Admissible traffic load of real time class of service for inter-domain
peers” in Proc. of ICAS/ICNS 2005, 23-28 October 2005, Papeete,
Tahiti, French Polynesia, published by IEEE Computer Society, 2005.
• The full text paper can be found at the homepage of TNT Group
http://tnt.tele.pw.edu.pl/include/members/Artikuly/Admissible.pdf