Bandwidth Management and
Scheduling in MPLS DiffServ
Networks
Ximing Zeng, Chung-Horng Lung,
Changcheng Huang
Department of Systems and
Computer Engineering
Carleton University
Ottawa, Canada
Anand Srinivasan
Eion
Ottawa, Canada
Outline
 Diverse QoS Requirements
 DiffServ Forwarding Classes
 Cisco Solutions
 Solutions based on fair queueing
 Our Solution
 Performance study
QoS requirements
Packet Loss
5%
Conversational
voice and video
0%
Zero
loss
100 msec
Command
/ control
(eg Telnet,
Interactive
games)
Voice/video
messaging
1 sec
Transactions
(eg E-commerce,
Web-browsing, Email access)
Streaming
audio/video
10 sec
Messaging,
Downloads
(eg FTP,
still image)
Delay
Fax
100 sec
Background
(eg Usenet)
QoS requirements for real time applications
Medium
Audio
Application
Conversational
voice
Degree of
symmetry
Two-way
Typical Data rates/
Amount of Data
4 - 64 kb/s
Key performance parameters and target values
End-to-end One
way Delay
Delay Variation
within a call
Information
Loss**
<150 msec
Preferred*
< 1 msec
< 3%
Packet
Loss Ratio
<400 msec limit*
Video
Videophone
Two-way
16 -384 kb/s
< 150 msec
preferred
<400 msec limit
Lip-synch :
< 100 msec
< 1%
Packet
Loss Ratio
Data
Telemetry
- two-way control
Two-way
<28.8 kb/s
< 250 msec
N.A
Zero
Data
Interactive
games
Two-way
< 1 KB
< 250 msec
N.A
Zero
Data
Telnet
Two-way
(asymmetric)
< 1 KB
< 250 msec
N.A
Zero
QoS requirements for interactive applications
Medium
Application
Degree of
symmetry
Typical data rate/
Amount of data
Key performance parameters and target values
One-way
Delay
(response time)
Delay
Variation
Information
loss
Audio
Voice
Messaging
Primarily
one-way
4-32 kb/s
< 1 sec for
playback
< 2 sec for
record
< 1 msec
< 3%
Packet
Loss Ratio
Data
Web-browsing
- HTML
Primarily
one-way
~ 10 kB
< 4 sec /page
N.A
Zero
Data
Transaction
services – high
priority
e.g. e-commerce,
ATM
Two-way
< 10 kB
< 4 sec
N.A
Zero
Data
E-mail
(server access)
Primarily
One-way
< 10 kB
< 4 sec
N.A
Zero
QoS requirements for streaming applications
Medium
Application
Degree of
symmetry
Data rate/
Amount of data
Key performance parameters and target values
Start-up
Delay
Transport delay
Variation
Packet loss at
session layer
Audio
Speech, mixed
speech and music,
medium and high
quality music
Primarily
one-way
5-128 kb/s
< 10 sec
< 1 msec
< 1% Packet loss
ratio
Video
Movie clips,
surveillance, realtime video
Primarily
one-way
16 -384 kb/s
< 10 sec
< 1 msec
< 1% Packet loss
ratio
Data
Bulk data
transfer/retrieval,
layout and
Synchronisation
information
Primarily
one-way
10 kB – 10 MB
< 10 sec
N.A
Zero
Data
Still image
Primarily
one-way
< 100 kB
< 10 sec
N.A
Zero
DiffServ Service Classes
 Expedited Forwarding (EF) PHB (RFC-2598)

Provides a low-loss, low-latency, low-jitter, and assured bandwidth
service. Real-time applications such as voice over IP (VoIP), video, and
online trading programs require such a robust network-treatment.
Assured Forwarding (AFxy) PHB (RFC-2597)
 Provides certain forwarding assurance by allocating certain bandwidth
and buffer space. Applications with certain QoS requirements but not realtime can use AF service. For example: streaming video.
Best Effort Service
No service guarantee except for a minimum bandwidth to prevent service
starvation.
Cisco Solution
 LLQ or MDRR
High priority EF
Low priority
AF and BE
EF
VoIP, Interactive Game…
AF1x
Video Conferencing…
AF2x
Video on demand …
AF3x
E-commerce …
AF4x
……
BE
http,ftp, email…
PQ
CBWFQ
Total reservable bandwidth is about 75%. BE reservation fixed around 25%.
EF traffic is constrained and should not exceed 33%; small queue and packet size.
AFs reserve the rest bandwidth.
Cisco Solution
BE = 25%
EF<33%
AF4
AF3
AF2
AF1
EF is assigned a bandwidth
less than 33% of the link speed
and is constrained according to
the assigned bandwidth.
However burst of EF traffic still
exists.
BE reserves a certain amount
of bandwidth.
The rest of the bandwidth can
be allocated to AF services.
AFs and BE may not always get their bandwidth as reserved !
Cisco Solution
 Advantage:EF packets are guaranteed smallest delay possible
by given them high priority.
 Tradeoff: AF packets may be delayed due to burst of EF packets
and cannot meet its desired delay bound!
bytes
bytes
ρ
delay
delay
backlog
backlog
σ
r
t
t
To minimize the impact EF brings to the AF classes:
EF has small queue size (therefore, close to CBR)
EF has small packet size (shorter waiting time for other packets)
Cisco Solution
 What if the EF traffic is bursty?
Link rate
Peak rate
Reserved rate
Average rate
EF traffic rate
t
AIftrade
bandwidth
off hasother
to bethan
made!
the The
average
peak
actual
raterate
of
bandwidth
EF
of EF
traffic
traffic
reserved
is claimed
is claimed
toallocatable,
EFallocatable,
class should
thenthen
AF
close
whento
QoS
isEF
guaranteed.
the
burst
peakcomes,
rateLow
tothe
minimize
bandwidth
bandwidth
theutilization!
service
to AF classes
impact.cannot be guaranteed. Bad
QoS!
Cisco Solution
EF average
BE=25%
EF wasted
AF4
AF3
AF1
AF2
Under LLQ, to minimize the
service impact to AF service
Classes, EF bandwidth is
Over-provisioned.
Other solutions
 Assign each class certain bandwidth
EF
VoIP, Interactive Game…
AF1x
Video Conferencing…
AF2x
Video on demand …
AF3x
E-commerce …
AF4x
……
BE
http,ftp, email…
WFQ/
DWRR
Other solutions (WFQ)
 Use weighted fair queueing to assign bandwidth to EF, AF and BE
classes.
 Advantage: Service to AF packets will not be affected by EF traffic, they
always get their reserved bandwidth
 Disadvantage: Over-provisioning is still needed to guarantee small
delay to EF classes.
r
delay
delay
σ
ρ
r If EF traffic is bursty, to have a small
delay, a large bandwidth reservation is
needed, which causes the same
problem of wasted bandwidth.
Other solutions (DWRR)
 Dynamically adjust the bandwidth to EF class according to the





backlog of EF traffic. The unused bandwidth can be used by BE
traffic.
Advantage: the EF still gets a relatively low delay.
Practical problems:
1)How often should we adjust the bandwidth allocation?
2)If it is WFQ, how can we adjust the virtual finish time for all the
backlogged packets on line?
3)The bandwidth unused by EF can be used by BE, but is there
any guaranteed minimum bandwidth? Or how can we assign the
unused bandwidth?
Our Solution
 Proposed scheduler architecture
High priority
EF
VoIP, Interactive Game…
Low priority
BE
http,ftp, email…
AF1x
Video Conferencing…
AF2x
Video on demand …
AF3x
E-commerce …
AF4x
……
PQ
CBWFQ
Our Solution
 EF and BE share the bandwidth:
EF
EF
r  max(ravg
 r BE , rres
)
Link rate
Peak rate
Reserved rate
rresEF
BE traffic rate
Average rate
EF traffic rate
t
EF
ravg
Our Solution
 EF and BE share the bandwidth:
EF
EF
r  max(ravg
 r BE , rres
)
Link rate
BE traffic rate
Peak rate
EF
r
Reserved rate res
Average rate r EF
avg
EF traffic rate
t
Our Solution
EF + BE =
EF
rBE + ravg
EF
ravg
rBE =
25%
EF
EF
- ravg
rres
AF4
AF3
AF1
AF2
The pie under WFQ
or Cisco LLQ
Now free!
AF4
AF3
AF1
AF2
The pie under our solution.
You can have an extra slice!
Our Solution
 Advantages:

EF is given no less (if not more) bandwidth than in WFQ.
Performance is guaranteed.
 AFs are guaranteed the same bandwidth, the same
performance can be expected.
 Bandwidth can be allocated to EF and AF users more efficiently!
 Tradeoff:

BE traffic may experience a longer delay due to EF bursts.
However, they are not delay sensitive and their average
minimum throughput is still guaranteed.
Simulation result
Src 0: EF traffic: 7 on-off voice sources
369 packets/sec in average.
Src 1: BE traffic: 800 packets/sce,
Exponential
Src 2/3: AF traffic: 400 packets/sce,
Exponential
Link speed: 2000 packets/sce.
Average load: 98.45%
Simulation result
 EF delay under LLQ, DWRR and WPRR
LLQ provides the smallest
Delay to EF class.
WPRR provides delay which is
Comparable tp LLQ
DWRR provides a much longer
Delay.
Simulation result
 AF delay under LLQ, DWRR and WPRR
The same delay bound is
Guaranteed under both DWRR
And WPRR.
Under LLQ, the AF delay is longer
Due to the burstness of EF traffic.
6% of the packets violate the delay
Bound.
Simulation result
 BE throughput
Simulation result
BE throughput
Conclusion
 We developed a new scheduler for DiffServ
routers with the following advantage:





High bandwidth utilization
Guaranteed QoS
Guarantee small delay and loss for EF.
Provide QoS guarantee to AF by reserving the bandwidth.
Guarantee the minimum throughput of BE.
References

[1] S.Blake, D.Black, M.Carlson, E.Davies, Z.Zhang, W.Weiss, “An Architecture for Differentiated Services.”
IETF RFC 2475. Dec 1998.

[2] V. Jacobson, K. Nichols, K. Poduri, “An Expedited Forwarding PHB.” IETF RFC 2598. June 1999.

[3] J. Heinanen, F. Baker, W. Weiss, J.Wroclawski, “Assured Forwarding PHB Group.” IETF RFC 2597. June
1999.

[4] J. Mao, W.M. Moh. B Wei, “PQWRR scheduling algorithm in supporting of DiffServ” 2001. ICC 2001. IEEE
International Conference on Communications,Volume: 3 , Pages:679 – 684.11-14 June 2001.

[5] A.K.Parekh, R.G.Gallager, “A Generalized Processor Sharing Approach to Flow Control in Integrated
Service Networks: The single node case,” IEEE/ACM Transactions on Networking, Pages:344 - 357. June 1993.

[6] H.Wang, C.Shen, K.G.Shin, “Adaptive-weighted packet scheduling for premium service” Communications,
2001. ICC 2001. IEEE International Conference on , Volume: 6 , Pages:1846 – 1850. 11-14 June 2001.

[7] F. Le Faucheur, L. Wu, B. Davie, S. Davari, P. Vaananen, R.Krishnan, P. Cheval, J. Heinanen, “Multi-Protocol
.
Label Switching (MPLS) Support of Differentiated Services.” IETF RFC 3270 May 2002
Thank You !