Traffic Engineering
Label Switched Paths
B i g o m o k e r o A n t o i n e
B a g u l a
T h e s i s p r e s e n t e d i n p a r t i a l f u l f i l m e n t o f t h e r e q u i r e m e n t s f o r t h e d e g r e e o f M a s t e r o f S c i e n c e a t t h e U n i v e r s i t y o f S t e l l e n b o s c h S u p e r v i s o r : P r o f . A . E . K r z e s i n s k i M a r c h 2 0 0 2Declaration
I,
the undersigned,
hereby declare that the work contained in this thesis is my own original work
and has not previously in its entirety or in part been submitted
at any university
for a degree .
SIgnature.:
iii
Abstract
T h e In te rn e t is e v o lv in g in to a c o m m e rc ia l p la tfo rm re q u irin g e n h a n c e d p ro to c o ls a n d a n e x p a n d e d p h y s ic a l in fra s tru c tu re a llo w in g a b e tte r d e liv e ry fro m IP . M u lti-p ro to c o l L a b e l S w itc h in g (M P L S ) is a te c h n o lo g y e n a b lin g tra ffic e n g in e e rin g a n d v irtu a l p riv a te n e tw o rk (V P N ) p ro v is io n in g . M P L S a c h ie v e s tra ffic e n g in e e rin g b y c a rry in g th e tra ffic o v e r v irtu a l c o n n e c tio n s c a lle d L a b e l S w itc h e d P a th s (L S P s ) w h ic h a re e n g in e e re d b a s e d o n Q o S re q u ire m e n ts s u c h a s d e la y , jitte r a n d p a c k e t lo s s m in im iz a tio n o r th ro u g h p u t m a x im iz a tio n .
T h is th e s is p ro p o s e s p a th fin d in g a n d tra ffic d is trib u tio n m e th o d s to b e d e p lo y e d in M P L S n e t-w o rk s fo r tra ffic e n g in e e rin g L S P s . A flo w o p tim iz a tio n m o d e l b a s e d o n a p re -p la n n e d ro u tin g a p -p ro a c h s e p a ra tin g p a th fin d in g a n d tra ffic d is trib u tio n is p re s e n te d . T h is m o d e l is a u g m e n te d b y a th re s h o ld ro u tin g a p p ro a c h w h ic h ro u te s th e tra ffic b a s e d o n th re s h o ld s e x p re s s in g th e m a x im u m lo a d le v e l re a c h e d b y n e tw o rk lin k s . T h is ro u tin g a p p ro a c h m o v e s th e tra ffic a w a y fro m th re s h o ld -m a rk e d lin k s to a c h ie v e lo w -u tiliz e d lin k s /p a th s . T h e p e rfo rm a n c e a n d ro u tin g c a p a b ilitie s o f th e s e m e th o d s a re e v a lu a te d th ro u g h d e s ig n e d s o ftw a re . A ro u tin g a rc h ite c tu re im p le m e n tin g a tw o -la y e r s ig n a llin g m o d e l fo r M P L S n e tw o rk is p ro p o s e d a n d e v a lu a te d th ro u g h s im u la tio n .
Opsomming
Die verandering
van die Internet
in 'n kommersiele
platform
met verbeterde
protokolle
en 'n
uitgebreide
fisieke infrastruktuur
stel die internetprotokol
(IP) in staat
tot beter lewering.
Multi-protokol-etiketskakeling
(MPLS),
is 'n tegnologie
vir die voorsiening
van televerkeerbeheer
en
virtuele
privaatnetwerke
(VPN).
MPLS
verskaf televerkeerbeheer
deur
die verkeer
te dra oar
virtuele
konneksies,
wat bekend staan
as etiketgeskakelde
paaie,
waarvan
die ontwerp
gebaseer
is op vereistes
vir diensgehalte
soos vertraging,
ritteling
en die minimering
van pakketverlies
of
maksimering
van deurvoer.
Hierdie tesis stel nuwe padvind- en verkeerdistribusiemetodes
voor wat aangewend
word in
MPLS-netwerke om etiketgeskakelde
paaie te beheer.
'n Model vir vloei-optimering-gebaseer
op
vooraf-beplande
roetering
wat padvinding
en verkeerdistribusie
skei-word
aangebied.
Hierdie
model
word uitgebrei
deur 'n benadering
van drempelroetering
wat die verkeer roeteer en gebaseer is op
drempels
wat die maksimum
ladingsvlak
voorstel wat bereik kan word deur netwerkskakels.
Hi-erdie roeteringsbenadering
skuif die verkeer weg van drempelgemerkte
skakels en bereik daardeur
laaggebruikte
skakelsjpaaie.
Die prestasie
en roeteringsvaardigheid
van hierdie
metodes
word
gevalueer deur selfontwikkelde
programmatuur.
'n Argitektuur
vir roetering
wat 'n
dubbellaag-seinmodel implementeer
vir 'n MPLS-netwerk,
word aangebied
en gevalueer met simulasie.
Acknowledgements
First my thanks
go to my advisor, Professor Anthony
E. Krzesinski
who by his sense of
responsi-bility, his insistence
on correctness
and precision, made this dissertation
possible.
I would like also to thank
my colleagues of the COE group for their
willingness
to help i many
occasions.
M y special thanks
go to Abriette
Senekal and Johan
de Kock whose contribution
to
this work is most valuable.
This work would not be possible without
the financial support
of TELKOM
SA and SIEM ENS
through
the Centre
of Excellence in ATM and Broadband
Networks
(COE).
M any friends, brothers,
and all my family have been supportive
in many ways since our arrival in
Cape Town. I would like to thank
all of them.
"Climbing
the mountain,
step by step, until we reach the top"
is the most important
principle
that
my wife Yvette,
my children Herman, Fortuna,
Nancy and Amani and myself learned
during
my seemingly never ending student
life. I would like to thank
deep from my heart
Yvette
for her
support
and love and the children for their love and friendship
by trying to be the best at school.
M entioning
each person
who has contributed
to the success of this
thesis
work is a hard
and
difficult task.
I,
therefore,
would like to apologize to those who are not explicitly
mentioned
here.
Contents
Abstract
v
Opsomming
vii
Acknowledgements
ix
1.4
Related
Work ...
1
Introduction
1.3
Traffic Regulation
Current
Status
of IP Networks
1
1
1
2
3
3
4
5
6
6
7
10
10
11
Key Contributions
Thesis Outline
. .
Traffic Regulation
QoS Architectures
Pricing the Internet
IP Traffic Management
Traffic Engineering
Internet
Protocols
Thesis Outline and Main Contributions
1.5.1
1.5.2
Traffic Engineering
1.4.1
1.4.2
1.1.1
1.1.2
1.1.3
1.1.4
1.5
1.2
1.1
2
Traffic
Engineering
Model
2.1
Introduction.
. . . .
13
13
2 .2 A B a s ic F lo w O p tim iz a tio n M o d e l
S o lu tio n C h a ra c te ris tic s . 2 .3
2 .2 .1
2 .2 .2
2 .2 .3
P a th Id e n tific a tio n . . . . .
P a th S e le c tio n a n d A c tiv a tio n .
T ra ffic D is trib u tio n .
1 4
15
1 6 1 825
2 .3 .1 W T D O p tim a lity . 2 5 2 .3 .2 F L D O p tim a lity . 2 7 C o n c lu s io n . P e rfo rm a n c e A n a ly s is 2 .4 2 .5 2 .3 .3 F a u lt-to le ra n c e28
29
35
3 .5 P e rfo rm a n c e A n a ly s is 3 T h r e s h o ld R o u tin g 3 .2 T h e T h re s h o ld R o u tin g A p p ro a c h . 3 .2 .1 L in k W e ig h t O p tim iz a tio n (L W O )3 .2 .2 D iffe re n tia te d R o u tin g R e g im e (D R R ) In tro d u c tio n . . . 3 7 3 7
38
38
40
48
48
48
51
51
5 353
5 8 T h e P a th -m u ltip le x e d M o d e l L in k C o s t O p tim iz a tio n (L C O ) P a th C h a ra c te ris tic s T h e P a th -d iffe re n tia te d M o d e l C o n c lu s io n . D iffe re n tia te d S e rv ic e s M o d e ls . . . . . 3 .4 .1 3 .4 .2 P e rfo rm a n c e M e a s u re m e n t . 3 .3 .1 3 .3 .2 3 .4 3 .6 3 .3 3 .1 4 P r o p o s e d R o u tin g A r c h ite c tu r e 4 .1 In tro d u c tio n . 4 .2 T h e P ro p o s e d R o u tin g A rc h ite c tu re61
6 162
CONTENTS
xiii
5
Traffic R egulation
4.3.5
Routing
Updates
4.4
Im plem entation
Strategy.
4.4.1
Traffic M easurem ent
5.2
The Traffic Regulation
M odel
5.2.1
Traffic Regulation
M ethods
5.2.2
Routing
Functions
5.3
Service
M odels
....
5.3.1
Equal
Services
5.3.2
Differentiated
Services
5.4
Fairness
M odels
.
5.4.1
The Traffic Regulation
Problem .
5.4.2
M ono-path
Fairness
Characteristics.
Introduction.
.
63
64
64
67
67
69
69
70
71
71
72
74
74
74
76
76
79
79
81
81
83
85
85
86
89
89
90
M ulti-path
Routing
Packet
Reordering
Traffic M easurem ent
Decentralized
Routing
4.4.2
The Edge-Edge
Control
Plane.
4.4.3
The User-User
Control
Plane
Conclusion
..
4.3.1
4.3.2
4.3.3
4.3.4
Scalability
Issues
. . . .
4.2.1
The Key Routing
Features
...
4.2.2
The Proposed
Network
Architecture
4.2.3
The Edge-Edge
Control
Plane.
4.2.4
The User-User
Control
Plane
4.2.5
The Proposed
Edge Router
Architecture.
5.1
4.5
4.3
5.4.3
M ulti-path
F airness C haracteristics.
5.5
S im ulation A nalysis
5.6
C onclusion
...
9294
98
6 C o n c lu sio n A p p e n d ix.1
M inim um P otential D elay
6.1
6.2 .2.3
T hesis S um m ary
F uture W ork
M ax-M in F airness
..
P roportional
F airness .
103
103
104
105
105106
107
List of Tables
2.1 L ink utilization distribution: 50 node netw ork
2.2 B ack-up paths .
2.3 lO a-node netw ork: path m ultiplicity
30
33
33
2.4 L ink utilization distribution: 50 node using F L D and K S P under m oderate load
(load factor 1.0) 34
2.5 L ink utilization distribution: 50 node netw ork using F L D and K S P under heavy
load (load factor
=
1.5) . . . .. 343.1 L W O vs P R E -P L A N N E D
3.2 D L I vs P R E -P L A N N E D
3.3 M IX d vs M IX p
3.4 D C R m ethod
3.5 P ath utilization
3.6 T raffic dropped: D L I m ethod
3.7 L ink utilization: lO a-node netw ork
3.8 P ath utilization: lO a-node netw ork
5.1 E xplicit routing.
5.2 E xplicit routing.
5.3 W T D using traffic halving.
5.4 W T D using a feedback factor 3/4.
5.5 W T D using different pow er values.
5.6 W T D using different queueing m odels.
54
55
55 56 5757
57 5896
96
97 97 97 97xv
List of Figures
2.1
2 .2 2 .32.4
2 .52.6
2.7
2 .8 3 .1 3 .2 3 .3 4 .1 4 .2 4 .34.4
5 .1 5 .2 5 .3 5 .4 5 .5 T h e "tra p " n e tw o rk T h e "e y e " n e tw o rk .W e ig h te d tra ffic d istrib u tio n u sin g d iffe re n t p o w e r v a lu e s .
T h e "sp o o n " n e tw o rk
T h e "fish " n e tw o rk . .
L in k u tiliz a tio n d istrib u tio n s
P a th u tiliz a tio n d istrib u tio n s
1 0 0 -n o d e n e tw o rk : n o rm a liz e d p a th le n g th d istrib u tio n s
T h e "lin k " th re sh o ld . . . .
T h e "m ix e d " c o st fu n c tio n
P a th u tiliz a tio n d istrib u tio n s
T h e "e d g e -e d g e " p la n e
T h e "L E R " ro u te r ..
T h e "le a rn in g " m o d e l
T h e "e d g e -e d g e fe e d b a c k " m o d e l
T h e "b o ttle n e c k " m o d e l . . . . .
S e rv ic e u n d e r d iffe re n t lig h t a n d h e a v y lo a d c o n d itio n s
T h e "lin e a r n e tw o rk " m o d e l . . th e "m u ltip a th fa irn e ss" m o d e l th e "te st n e tw o rk " m o d e l ... x v ii
15
25
26
2 728
3 0 3 0 3 1 4 2 5 0 5 6 6 568
7 3 7 586
87
90
94
9 55.6
DropTail queueing model
5.7
RED queueing model.
5.8
SFQ queueing model .
99
100
Chapter
1
Introduction
1.1
Current Status of IP Networks
C o m m u n ic a tio n n e tw o r k s h a v e e x p e r ie n c e d a r e v o lu tio n d u r in g th e la s t d e c a d e r e s u ltin g f r o m r a p id d e v e lo p m e n ts in h a r d w a r e a n d s o f tw a r e w h ic h m a d e a m ix tu r e o f n a r r o w - a n d w id e - b a n d a p p lic a tio n s a f f o r d a b le to a m u c h la r g e r g r o u p o f u s e r s . F o llo w in g th is r e v o lu tio n , th e I n te r n e t is e v o lv in g in to a c o m m e r c ia l p la tf o r m r e q u ir in g m o r e th a n th e b e s t- e f f o r t s e r v ic e s u p p o r te d b y c u r r e n t I n te r n e t p r o to c o ls , a n d a n e x p a n d e d p h y s ic a l in f r a s tr u c tu r e a llo w in g a b e tte r d e liv e r y s e r v ic e f r o m I P . A s a r e s u lt, th e I n te r n e t b a c k b o n e is b e in g u p g r a d e d in to a c o r e o p tic a l tr a n s p o r t a n d m a n y e f f o r ts a r e d e p lo y e d to m a k e m o r e e f f ic ie n t u s e o f
IP
b a c k b o n e p r o to c o ls b y h a v in g th e s e p r o to c o ls r u n n in g d ir e c tly o n to p o f o p tic a l c r o s s - c o n n e c ts . S e v e r a l o th e r s o lu tio n s a r e b e in g in v e s tig a te d b y th eIP
c o m m u n ity to d e liv e r b e tte rIP
s e r v ic e s in a s c a la b le f a s h io n . T h e s e in c lu d e th e m o v e f r o m a r o u tin g to a s w itc h in g te c h n o lo g y w ith in th e I n te r n e t b a c k b o n e , th e in te g r a tio n o f e n h a n c e d p r o to c o ls a n d Q o S a r c h ite c tu r e s in th e n e tw o r k a n d th e d e v e lo p m e n t o f n e w m e c h a n is m s f o r r e s o u r c e p r ic in g a n d tr a f f ic m a n a g e m e n t.1.1.1
Internet Protocols
N e ith e r th e I P v 4 r o u tin g p r o to c o ls c u r r e n tly u s e d in th e I n te r n e t n o r th e e m e r g in g I P v 6 r o u tin g p r o to c o ls p r o v id e th e Q o S r e q u ir e d f o r r e a l- tim e a n d b a n d w id th in te n s iv e a p p lic a tio n s in th e m o d e r n I n te r n e t. C u r r e n t g e n e r a tio n I P v 4 p r o to c o ls w e r e d e v e lo p e d o n th e b a s is o f a c o n n e c tio n -le s s m o d e l w h e r e e a c h p a c k e t's n e x t h o p a n d o u tp u t p o r t a r e d e te r m in e d b y a c o m p u ta tio n a lly e x p e n s iv e lo n g e s t- p r e f ix - m a tc h m e c h a n is m a n d r o u tin g d e c is io n s a r e b a s e d o n s im p le m e tr ic s s u c h a s d e la y o r h o p - c o u n t w h ic h le a d to th e s e le c tio n o f s h o r te s t- p a th r o u te s . D e s p ite its a b ility to s c a le to v e r y la r g e n e tw o r k s , th is a p p r o a c h p r o v id e s o n ly a r u d im e n ta r y Q o S w h ic h d o e s n o t m e e t th e d e m a n d f o r s c a la b ility r e q u ir e d f o r b a n d w id th in te n s iv e a p p lic a tio n s in m o d e r n n e tw o r k s . D e s p ite its im p r o v e m e n t o v e r th e I P v 4 p r o to c o ls b y p r o v id in g im p r o v e d p r io r ity s c h e m e s , f lo w la b e l f ie ld s
a n d e x te n d e d h e a d e rs, th e e m e rg in g g e n e ra tio n IP v 6 p ro to c o ls p ro v id e v e ry fe w Q o S fe a tu re s c o m p a re d to th e n e e d fo r Q o S in re a l-tim e a p p lic a tio n s. T h e d e sig n o f Q o S ro u tin g p ro to c o ls is th e re fo re a n im p o rta n t issu e u p o n w h ic h th e p e rfo rm a n c e o f th e m o d e rn In te rn e t w ill d e p e n d .
1.1.2
QoS Architectures
Q o S g e n e ra lly re fe rs to v a rio u s m e c h a n ism s a im in g a t p ro v id in g p re fe re n tia l tre a tm e n t to c e rta in ty p e s o f tra ffic w ith th e o b je c tiv e o f im p ro v e d n e tw o rk p e rfo rm a n c e (in c re a se d th ro u g h p u t, d e la y a n d jitte r m in im iz a tio n , p a c k e t lo ss re d u c tio n , e tc .). S e v e ra l so lu tio n s fro m th e In te rn e t E n g in e e r-in g T a sk F o rc e (IE T F ) a n d th e re se a rc h c o m m u n ity h a v e b e e n p ro p o se d to su p p o rt Q o S in th e In te rn e t. T h e IE T F e ffo rts h a v e b e e n fo c u se d o n th e d e fin itio n o f sta n d a rd s fo r Q o S m e c h a n ism s w ith in th e n e tw o rk w h ile th e a c a d e m ic , v e n d o r a n d o p e ra to r c o m m u n itie s h a v e b e e n m o re in v o lv e d in o p tim iz a tio n te c h n iq u e s (o p tim a l ro u tin g a n d o p tim a l re so u rc e m a n a g e m e n t) to b e d e p lo y e d in n e tw o rk e n v iro n m e n ts su p p o rtin g IE T F Q o S sta n d a rd s.
T w o Q o S a rc h ite c tu re s h a v e b e e n sta n d a rd iz e d b y th e IE T F fo r Q o S su p p o rt in th e In te rn e t: In -te g ra -te d S e rv ic e s (In tse rv ) a n d D iffe re n tia te d S e rv ic e s (D iffse rv ) a rc h ite c tu re s (In tse rv ) (D iffse rv ). T h e In tse rv is a Q o S a rc h ite c tu re th a t o ffe rs a re lia b le se rv ic e lik e th e c irc u it-sw itc h e d te le p h o n e n e tw o rk b y m a in ta in in g a p e r-flo w sta te w ith in th e n e tw o rk . D iffse rv is a sta te le ss a rc h ite c tu re b a se d o n th e a g g re g a te d se rv ic e c o n c e p t su p p o rtin g d iffe re n t se rv ic e le v e ls a n d a llo w in g th e se d if-fe re n t se rv ic e le v e ls to b e tre a te d d iffe re n tly a t th e in g re ss o f th e n e tw o rk . V a rio u s m e c h a n ism s h a v e b e e n p ro p o se d w ith in th e fra m e w o rk s o ffe re d b y th e In tse rv a n d D iffse rv a rc h ite c tu re s fo r a d d in g se rv ic e -v a lu e s to th e b e st-e ffo rt se rv ic e in o rd e r to su p p o rt Q o S in th e In te rn e t. T h e se in c lu d e :
• th e a c tiv a tio n o f n e w se rv ic e s a n d re so u rc e a llo c a tio n p o lic ie s to m e e t c u sto m e r n e e d s,
• th e d e p lo y m e n t o f c o n g e stio n a v o id a n c e , tra ffic c la ssific a tio n a n d p rio ritiz a tio n m e c h a n ism s a llo w in g a n a p p ro p ria te n e tw o rk h a n d lin g to m e e t c u sto m e r re q u ire m e n ts a n d w illin g n e ss to p a y fo r se rv ic e s,
• n e tw o rk m e te rin g to su p p o rt b illin g /a c c o u n tin g ,
• n e tw o rk m a n a g e m e n t a n d n e tw o rk m o n ito rin g a n d e ffic ie n t b a n d w id th m a n a g e m e n t to e n -fo rc e b a n d w id th c o m m itm e n ts to tra ffic so u rc e s, a n d
• th e im p le m e n ta tio n o f a fle x ib le p a c k e t h a n d lin g p o lic y to a v o id b a n d w id th a llo c a tio n s th a t e x c e e d th e a v a ila b le re so u rc e s.
T h e se m e c h a n ism s h a v e b e e n im p le m e n te d in th e In te rn e t u sin g v a rio u s to o ls in v o lv in g (1 ) q u e u e -in g m e c h a n ism s su c h a s F irst-In -F irst O u t (F IF O ), P rio rity Q u e u e in g (P Q ), C la sse d -b a se d Q u e u e -in g (C B Q ), W e ig h te d F a ir Q u e u e in g (W F Q ), a c tiv e q u e u e in g m e c h a n ism s su c h a s R a n d o m E a rly D e te c tio n (R E D ) a n d (2 ) tra ffic sh a p in g a n d p o lic in g u sin g le a k y -b u c k e t a n d to k e n -b u c k e t m e c h -a n ism s a n d Q o S sig n a llin g u sin g D iffse rv
IP
p re c e d e n c e o r R S V P m e c h a n ism s.1.1 Current Status of IP Networks
1.1.3
Pricing the Internet
3
T he conversion of the Internet into a com m ercial platform raises the problem of resource pricing
w ithin the Internet. It is know n for exam ple that past V P N provisioning efforts deployed by the
IP
com m unity failed due to the lack of an effective netw ork infrastructure and resource pricing m
ech-anism s that should be im plem ented by the Internet service providers (IS P s) to recover the costs of
the investm ents m ade in the Internet infrastructure to support V P N provisioning. A n em erging
best-effort architecture (S airam esh et al., 1995) (O dlyzko, 1999) (K elly et al., 1998) (M
ackie-M ason and V arian, 1995) based on pricing m echanism s is being extensively investigated by the
IP
com m unity. It is based on a single class m odel allow ing users to m ark packets based on thecost of forw arding packets w ithin the netw ork and uses feedback m echanism s to allow users to
adjust their transm ission rates according to the price of the resources. U sing this m odel,
differ-entiation of services w ill be based on a m arket-pricing m odel w here resources w ill be priced by
the netw ork based on the congestion caused by the traffic in the netw ork, and users w ill im
ple-m ent a w illingness-to-pay schem e w here resources w ill be purchased according to the im portance
(priority) accorded to the traffic. T his m odel raises the expectation that the m odern Internet w ill
still be dom inated by best-effort traffic but w ith m ore control to support congestion avoidance
and differentiation of services. H ow pricing and different congestion avoidance m echanism s w ill
be im plem ented in the best-effort subnetw ork is an issue w hich is still under investigation.
1.1.4
IP Traffic Management
R eal-tim e traffic m anagem ent has been successfully im plem ented in circuit-sw itched netw orks
to im prove the grade of service offered by a netw ork by ensuring that the netw ork utilization
is m axim ized under all traffic profiles, including long-term , short-term and interm ediate-term
variations of the traffic. A s deployed in circuit sw itching netw orks, real-tim e traffic m anagem ent
provides m onitoring of netw ork perform ance through a set of control m echanism s. T he transition
of the Internet from a non-cooperative netw ork into a cooperative environm ent requires that
sim ilar control m echanism s to those deployed in real-tim e traffic m anagem ent of circuit-sw itched
netw orks be im plem ented in
IP
netw orks for Q oS support. H ow ever, despite their successfuldeploym ent in telephone netw orks, real-tim e traffic m anagem ent m echanism s have been either
scarcely im plem ented in
IP
netw orks or scarcely find their counter-parts inIP
netw orks. Further-m ore, the traffic urther-m anageurther-m ent m echanism s currently deployed in
IP
netw orks com bining routingand resource m anagem ent approaches to optim ize system -w ide m easures of perform ance such
as average response tim e, throughput, delay, etc. are inappropriate in the m odern Internet:
these approaches do not consider the com plexity of the m odern Internet characterized by the
continually increasing size (num ber of system s and users) and heterogeneity of protocols (T C P ,
U D P , P P P ), of applications (telnet, F T P , W E B ) and resources (C P U , m em ory, bandw idth,
1.2
Traffic Engineering
T ra ffic e n g in e e rin g is a n im p o rta n t a s p e c t o f n e tw o rk m a n a g e m e n t w h ic h is e x p e c te d to b e e x -te n s iv e ly im p le m e n te d in e m e rg in g In te rn e t p ro to c o ls fo r re a l-tim e tra ffic m a n a g e m e n t s u p p o rt. T ra ffic e n g in e e rin g a llo w s d a ta to b e e ffic ie n tly ro u te d th ro u g h th e n e tw o rk b y im p le m e n tin g Q o S a g re e m e n ts b e tw e e n th e a v a ila b le re s o u rc e s a n d th e c u rre n t a n d e x p e c te d tra ffic . M u lti-p ro to c o l la b e l s w itc h in g (M P L S ) (D a v ie e t a l., 1 9 9 8 ) h a s b e e n p ro p o s e d b y th e IE T F to e x te n d th e IP v 4 d e s tin a tio n -b a s e d ro u tin g p ro to c o ls to p ro v id e n e w a n d s c a la b le ro u tin g c a p a b ilitie s in c lu d in g tra ffic e n g in e e rin g a n d tra ffic flo w d iffe re n tia tio n b a s e d o n e x p lic it ro u tin g . M P L S is a s w itc h -in g te c h n o lo g y u s e d b y s e rv ic e p ro v id e rs a n d e n te rp ris e s to p ro v id e c o n n e c tio n -o rie n te d s e rv ic e s o v e r c o n n e c tio n -le s s n e tw o rk in fra s tru c tu re s . M P L S h a s b e e n a d o p te d b y m a n y In te rn e t s e rv ic e p ro v id e rs (IS P s ) to b e u s e d in th e ir
IP
W A N n e tw o rk fo r tw o m a in re a s o n s : th e n e e d to re p la c e th e c o m p le x a n d e x p e n s iv e o v e rla y a rc h ite c tu re b y a m o re s c a la b le a rc h ite c tu re , a n d th e n e e d to p ro v id e c la s s o f s e rv ic e (C O S ), tra ffic e n g in e e rin g a n d v irtu a l p riv a te n e tw o rk s (V P N s ) a t th eIP
la y e r. T h e o v e rla y m o d e l w a s in tro d u c e d a s a m e a n s o f u p g ra d in g th e In te rn e t b a c k b o n e b y a llo w in g a nIP
fo rw a rd in g n e tw o rk to b e o v e rla y e d u p o n a s w itc h e d p h y s ic a l n e tw o rk c o m p o s e d o f A T M s w itc h e s . T h is m o d e l re q u ire s a m e s h e d d e s ig n w h e re th e n u m b e r o f la y e r-2 c irc u its n e e d e d fo r th e in te r-c o n n e c tio n o fIP
ro u te rs in c re a s e s e x p o n e n tia lly w ith th e n u m b e r o f ro u te rs to b e in te rc o n n e c te d . T h is re s u lts in s c a la b ility is s u e s w h ic h a re s o lv e d b y th e M P L S te c h n o lo g y b y m a k in g W A N s w itc h e sIP
v is ib le . B y s e p a ra tin g th e ro u tin g (c o n tro l fu n c tio n s s u c h a s ro u te lo o k u p ) fro m th e fo rw a rd in g o fIP
p a c k e ts , M P L S h a s s o lv e d th e p ro b le m s re la te d to th e e x p e n s iv e lo n g e s t-p re fix m a tc h u s e d in c u rre n tIP
n e tw o rk s a n d th e n e e d fo r c o m p lic a te d n e tw o rk a d d re s s tra n s la tio n (N A T ) re q u ire d to tra n s la te ille g a l, p riv a te o r d u p lic a te a d d re s s e s . F u rth e rm o re , th e a b ility to e a s ily m a p c irc u it-b a s e d la y e r-2 Q o S o n to la y e r-3 is a n o th e r fe a tu re th a t m a k e s M P L S a n id e a l p ro to c o l fo r th e In te rn e t b a c k b o n e .M P L S w a s in itia lly p ro p o s e d b y th e In te rn e t E n g in e e rin g T a s k F o rc e (IE T F ) a s a to p o lo g y -d riv e n ro u tin g m o d e l w h ic h d e s p ite its re la tiv e im p ro v e m e n t o v e r c o n v e n tio n a l
IP
ro u tin g b y p ro v id in g h ig h e r p ro c e s s in g s p e e d s a n d d iffe re n tia tio n o f tra ffic flo w s , is c o n s tra in e d b y th e s a m e s h o rte s t-p a th s e le c tio n th a t a p p lie s in c o n v e n tio n a lIP
n e tw o rk s . C o n s tra in t-b a s e d ro u tin g (J a m o u s s i, 1 9 9 9 ) w a s n e x t p ro p o s e d a s a n im p ro v e m e n t o v e r th e b a s ic to p o lo g y -d riv e n M P L S to s u p p o rt tra ffic e n g in e e rin g b y a llo w in g p a c k e ts b e lo n g in g to a flo w to b e s e n t to th e d e s tin a tio n u s in g a rb itra ry v irtu a l p a th s (s h o rte s t a n d n o n -s h o rte s t) w h ic h h a v e b e e n e n g in e e re d b a s e d o n th e flo w 's c o n s tra in ts (b a n d w id th g u a ra n te e , la te n c y a n d jitte r m in im iz a tio n o r s o m e o th e r p e rfo rm a n c e re q u ire m e n ts ). T h e s e v irtu a l p a th s a re c a lle d c o n s tra in t-ro u te d L S P s (C R -L S P s ). In a d d itio n to its c o n s tra in t-ro u tin g s u p p o rt, M P L S a ls o p ro v id e s a n im p ro v e d fo rm o f e x p lic it ro u tin g c o m p a re d to c o n v e n tio n a l d a ta g ra m n e tw o rk s . T ra d itio n a l d a ta g ra m n e tw o rk s p ro v id e a fo rm o f e x p lic it ro u tin g w h e re th e n e tw o rk -la y e r a d d re s s o f e a c h n o d e a lo n g th e e x p lic it p a th is a tta c h e d to e a c h p a c k e t. T h is re s u lts in la rg e o v e rh e a d s in th e p a c k e t h e a d e r. In c o n tra s t, M P L S a llo w s a le s s e x p e n s iv e fo rm o f e x p lic it ro u tin g w h e re la b e l s w a p p in g is u s e d to id e n tify a n L S P re g a rd le s s o f w h e th e r th e L S P is e s ta b lis h e d b y h o p -b y -h o p o r e x p lic it ro u tin g .1.3 Traffic Regulation
1.3
Traffic Regulation
5
The evolution of the Internet from a best-effort netw ork into a com m ercial platform requiring
resource pricing leads to a trade-off betw een the econom ic ram ifications of the bandw idth sharing
policy deployed in the netw ork and the engineering efficiency achieved by the traffic regulation
algorithm s im plem ented. N etw ork cooperation through Q oS signalling is at the heart of traffic
regulation m echanism s. N etw ork cooperation allow s the edge of the netw ork to cooperate w ith its
core to im prove the perform ance of the system , to m inim ize the im pact of congestion and provide
end-to-end service levels and Q oS policies to m eet custom er requirem ents.
Q oS signalling is a form of netw ork com m unication w hich allow s an end-application to signal
its bandw idth requirem ents and Q oS requirem ents to the netw ork by conveying the inform ation
in the packet header as in the
IP
precedence schem es used for differentiated Q oS or by using asignalling protocol such as R SV P. O ptim ality through the use of feedback m echanism s and Q oS
support through the use of fair rate control schem es are the tw o m ost im portant aspects of traffic
regulation. H ow different transm ission rates are allocated to different paths in the best effort
sub-netw ork is therefore an im portant aspect that affects the perform ance of traffic regulation schem es
and the Q oS achieved by the netw ork. Feedback m echanism s have been w idely used in A TM for
congestion control of A B R traffic. The EC N protocol w as extensively investigated by the Internet
com m unity to support feedback m echanism s allow ing applications to adjust their transm ission
rate to the observed netw ork load. Feedback schem es are classified into tw o categories referred to
as explicit and im plicit feedback schem es. Im plicit feedback also called bit-based feedback schem es
are based on a binary indication of congestion issued by the netw ork to allow users to adapt their
transm ission rate to an estim ation of the netw ork load. These schem es require m inim al processing
from the netw ork: m inim al participation is required from core routers and sw itches and m inim al
exchange of inform ation betw een the netw ork and users/applications. Explicit feedback schem es
involve a distributed com putation of rate allocations w here the transm ission rates are com puted
by the netw ork and sent to users/applications as feedback inform ation to adapt their transm
is-sion rate to an estim ation of the netw ork load. Though m oving the processing overhead tow ard
the netw ork, this schem e achieves perform ance properties such as efficiency, fairness, controlled
queueing delay and robustness. The integration of feedback schem es based on feedback m
echa-nism s in the existing and em erging routing architectures raises an issue related to the localization
of the processing entities in the routing environm ent and the nature of the signalling protocols
1.4
Related Work
1.4.1
Traffic Engineering
The
traffic engineering
problem
may be addressed
using either
a traffic distribution
approach
consisting
of an optimal
mapping
of the offered traffic to the available resources
(flow routing)
or
a capacity
engineering
approach
(also called capacity
routing)
consisting
of sizing the network to
optimize network resources.
W hile traffic distribution
relates to flow allocation
problems,
capacity
engineering
is more concerned
with capacity
allocation
problems.
Traffic Distribution Approaches
Several
traffic
distribution
proposals
based
either
on a shortest-path
or a multi-path
routing
scheme for adding
traffic engineering
capabilities
to the best-effort
Internet
were proposed
by
IP
researchers .
• Shortest-path
routing
schemes:
Shortest-path
routing
schemes achieve traffic engineering
by modification
of link metrics to adjust the mapping
of the traffic to resources.
The routing
model
presented
in (Fortz
and Thorup,
2000) consisting
of engineering
the
IP
traffic by
optimizing
OSPF
weights is a good illustration
of the application
of shortest-path
routing
in
IP
traffic engineering.
Despite
their
relative
simplicity
and the advantage
of being piggy-backed
on existing
pro-tocols, these approaches
are well suited to off-line environments
where traffic demands
are
known in advance and may lead to random
traffic shifts which, if uncontrolled,
may lead to
performance
degradations
.
• Multi-path
routing
schemes:
M ulti-path
traffic engineering
approaches
consist of the
com-putation
of several
paths
for routing
the traffic offered to a source-destination
pair
and
load-balancing
these paths to increase the network performance.
The requirements
for traffic engineering in M PLS networks are presented
in (Awduche et al.,
1998).
A first attempt
to implement
multi-path
routing
in the Internet
was made
in the OSPF
Equal Cost M ulti-path
protocol (ECM ) which consisted of load-balancing
the traffic between
several paths
by halving
the traffic carried
by a congested
path
and re-routing
half of this
traffic over an alternate
path.
This scheme is based on an unbalanced
traffic distribution
which may lead to oscillation
by continually
shifting the overload from a congested
path to
an uncongested
path.
Optimized
M ulti-path
and M PLS M ulti-path
were later proposed
by Villamizar
(Villamizar,
1999) to correct the unbalanced
distribution
of traffic experienced
by the ECM protocol and
avoid the related
problem
of oscillation.
1 .4 R e la te d W o r k
7
W id ja ja (W id ja ja , 1 9 9 8 ) p ro p o s e d M A T E , a tra ffic e n g in e e rin g s c h e m e b a s e d o n p e rio d ic a lly p ro b in g m u ltip le p a th s a n d re d is trib u tin g th e tra ffic in o rd e r to b a la n c e lo a d s .
C a p a c ity
E n g in e e r in g
A p p r o a c h e s
P ro p o s a ls fo r b a n d w id th a llo c a tio n re fe rre d to a s c a p a c ity e n g in e e rin g (L a i, 2 0 0 0 ) a n d s e v e ra l a p p ro a c h e s fo r s h a rin g th e a v a ila b le b a n d w id th o f a n e tw o rk b e tw e e n v irtu a l n e tw o rk s h a v e b e e n p re s e n te d b y th e In te rn e t E n g in e e rin g T a s k F o rc e (IE T F ) (A s h , 2 0 0 1 ).
C a p a c ity ro u tin g a p p ro a c h e s to tra ffic e n g in e e rin g h a v e b e e n e x te n s iv e ly s tu d ie d u s in g n o n -lin e a r a n d lin e a r m o d e ls to m a x im iz e a re w a rd fu n c tio n o r m in im iz e a p e n a lty fu n c tio n e x p re s s in g th e p e rfo rm a n c e re q u ire d fro m th e n e tw o rk . M a n y h e u ris tic s fo r s o lv in g c a p a c ity ro u tin g p ro b le m s h a v e b e e n p ro p o s e d in th e lite ra tu re :
• M E N T O R w a s p re s e n te d in (K e rs h e n b a u m e t a l., 1 9 9 1 ) a s a n h e u ris tic a lg o rith m s o lv in g a m in im u m c o s t to p o lo g y c o m p u ta tio n p ro b le m fo r a m e s h n e tw o rk a p p lic a b le to th e p ro b le m o f o b ta in in g s ta rtin g to p o lo g ie s fo r o th e r n e tw o rk d e s ig n p ro c e d u re s .
• F lo w d e v ia tio n (F ra tta e t a l., 1 9 7 3 ) a n d p ro x im a l d e c o m p o s itio n (C h ifH e t e t a l., 1 9 9 4 ) w e re p ro p o s e d a s h e u ris tic m e th o d s s o lv in g th e lo a d in g p ro b le m b y u s in g th e a v e ra g e p a c k e t d e la y a s m in im iz a tio n o b je c tiv e .
• In (W a n g a n d W a n g , 1 9 9 9 ), th e c a p a c ity ro u tin g p ro b le m is fo rm u la te d a s a n o p tim iz a tio n p ro b le m w ith a c a p a c ity s c a le fa c to r in th e o b je c tiv e fu n c tio n .
• T h e re -ro u tin g h e u ris tic a lg o rith m p ro p o s e d in (B a ra h o m a , 1 9 9 6 ) s o lv e s th e c la s s ic a l n e tw o rk lo a d in g p ro b le m u s in g d is c re te lin k c a p a c ity a s s ig n m e n ts .
• A c o s t m in im iz a tio n lo a d in g p ro b le m a p p lie d to IP n e tw o rk s w h ic h e x p lic itly c o n s id e rs th e re s tric tio n s im p o s e d b y th e O S P F ro u tin g p ro to c o l is p re s e n te d in (B e n m o h a m e d e t a l., 1 9 9 8 ) .
• In (K o d ia la m a n d L a k s h m a n , 2 0 0 0 ) K o d ia la m a n d L a k s h m a n p re s e n t a n a lg o rith m fo r th e d y n a m ic ro u tin g o f b a n d w id th g u a ra n te e d tu n n e ls in o n -lin e e n v iro n m e n ts b a s e d o n a m in -im u m in te rfe re n c e ro u tin g c o n c e p t w h ic h u s e s th e id e a th a t a n e w ly ro u te d tu n n e l m u s t fo llo w a ro u te th a t d o e s n o t in te rfe re to o m u c h w ith a ro u te th a t m a y b e c ritic a l to s a tis fy a fu tu re d e m a n d .
1 .4 .2
T r a f f ic R e g u la tio n
T ra ffic re g u la tio n o f d a ta g ra m tra ffic h a s b e e n im p le m e n te d fo llo w in g th re e d iffe re n t a p p ro a c h e s . T h e s e in c lu d e a p p ro a c h e s d e p lo y e d fo r A T M A B R tra ffic , n o n -p ric in g a d a p tiv e ra te a lg o rith m s a n d th e e m e rg in g IP p ric in g m e c h a n is m s im p le m e n te d in th e c o n g e s tio n c o n tro l o f th e T C P p ro to c o l s ta c k .
Non-pricing
Algorithms
T r a f f ic r e g u la tio n is s tr o n g ly r e la te d to th e c o n c e p t o f n e tw o r k f a ir n e s s a n d th e a d a p tiv e b a n d w id th s h a r in g a lg o r ith m s u s e d to s h a r e th e n e tw o r k b a n d w id th b e tw e e n c o m p e titiv e f lo w s . N e tw o r k f a ir n e s s is r e la te d to h o w th e lin k b a n d w id th is s h a r e d b e tw e e n s h o r te s t a n d lo n g e s t p a th s . F o u r f a ir n e s s c r ite r ia h a v e b e e n p r e s e n te d in th e lite r a tu r e : M a x im u m th r o u g h p u t f a ir n e s s , M a x - m in f a ir n e s s , P r o p o r tio n a l f a ir n e s s , M in im u m p o te n tia l d e la y f a ir n e s s a n d W e ig h te d s h a r e s f a ir n e s s .• A n e x p lic it a lg o r ith m c o n v e r g in g in a f in ite n u m b e r o f ite r a tio n s to a n e x a c t m a x - m in f a ir r a te a llo c a tio n is p r e s e n te d in ( C h a r n y e t a l., 1 9 9 5 ) . I t is b a s e d o n a d is tr ib u te d s c h e m e w h e r e u s e r s p r o g r e s s iv e ly d is c o v e r th e ir r a te a llo c a tio n b y c o m p a r is o n w ith th e a d v e r tis e d r a te o f th e lin k s o n its r o u te .
• T h e s tu d y p r e s e n te d in ( M a s s o u lie a n d J .R o b e r ts , 2 0 0 0 ) b y M a s s o u lie a n d R o b e r ts is b a s e d o n a n a lte r n a tiv e a p p r o a c h a v o id in g th e c o m p le x ity o f e x p lic it r a te c a lc u la tio n s b y im p le m e n tin g e ith e r a n e n d - to - e n d w in d o w c o n tr o l o r r a te a d ju s tm e n ts p e r f o r m e d b y u s e r s in r e s p o n s e to b in a r y c o n g e s tio n s ig n a ls to d e r iv e a lg o r ith m s m im ic k in g th e f o u r d if f e r e n t f a ir n e s s c r ite r ia .
• A m o r e g e n e r a l a d a p tiv e r a te a lg o r ith m in c lu d in g th e f o u r f a ir n e s s c r ite r ia p r e s e n te d a b o v e w a s s tu d ie d b y M o a n d W a lr a n d in ( M o a n d W a lr a n d , 2 0 0 0 ) .
• J a in a n d C h iu ( J a in a n d C h iu , 1 9 8 9 ) s tu d ie d th e im p a c t o f v a r io u s f e e d b a c k c o n tr o l m o d e ls o n th e s h a r in g o f a s in g le lin k b e tw e e n c o m p e titiv e f lo w s . T h e ir w o r k r e s u lte d in th e A d d itiv e I n c r e a s e /M u ltip lic a tiv e D e c r e a s e ( A I M D ) a lg o r ith m d e p lo y e d in th e c o n g e s tio n c o n tr o l o f th e T C P p r o to c o l.
Market-pricing
Algorithms
I P p r ic in g m e c h a n is m s a r e s u b d iv id e d in to th r e e m o d e ls r e f e r r e d to a s f la t- p r ic in g , u s a g e - s e n s itiv e p r ic in g a n d c o n g e s tio n - s e n s itiv e . 1 .Flat-pricing:
P r ic in g o f I n te r n e t a c c e s s is c u r r e n tly d o m in a te d b y a f la t- p r ic in g s tr u c tu r e b a s e d o n a n e q u a l s e r v ic e m o d e l w h e r e a f ix e d m o n th ly f e e w a s p a id b y u s e r s to a c c e s s th e lin k to th e n e tw o r k . W h ile b e in g q u e s tio n e d b y s e v e r a l r e s e a r c h e r s , th e im p le m e n ta tio n o f a f la t-p r ic in g is s till b e lie v e d to b e a n a c c e p ta b le c h o ic e f o r p r ic in g th e m o d e r n I n te r n e t ( O d ly z k o , 1 9 9 9 ) . T h is b e lie f is b a s e d o n th e o p in io n th a t th e c u r r e n t I n te r n e t is o v e r - e n g in e e r e d to a c c o m m o d a te e v e n p e a k lo a d s a n d lik e ly to r e m a in th a t w a y .2. Usage-sensitive
pricing:
T h e d e p lo y m e n t o f a u s a g e - s e n s itiv e p r ic in g s tr u c tu r e is e m e r g in g w h e r e r e s o u r c e s a r e p r ic e d a c c o r d in g to th e Q o S r e q u ir e d b y I n te r n e t a p p lic a tio n s . T h is m e c h a n is m is b a s e d o n th e o p in io n th a t s c a r c e r e s o u r c e s s u c h a s b a n d w id th a n d b u f f e r s p a c e m u s t b e p r ic e d d if f e r e n tly a c c o r d in g to th e g r a d e o f s e r v ic e ( G o S ) r e c e iv e d b y th e s e a p p lic a tio n s to p r o m o te a r a tio n a l u s e o f r e s o u r c e s . S e v e r a l s tu d ie s a d d r e s s in g r e s o u r c e m a n a g e m e n t u s in g a u s a g e - s e n s itiv e p r ic in g s c h e m e s h a v e b e e n in v e s tig a te d in th e lite r a tu r e .1 .4 R e l a t e d W o r k
9
• Q o S P r i c i n g M o d e l
Q o S p r ic in g is p r e s e n te d in ( S a ir a m e s h e t a l., 1 9 9 5 ) a n d ( S a ir a m e s h , 1 9 9 7 ) a s a n a p -p r o a c h to -p r ic in g , o -p tim a l r e s o u r c e a llo c a tio n a n d Q o S p r o v is io n in g in h ig h - s p e e d p a c k e t n e tw o r k s . I t is b a s e d o n a n e c o n o m ic m o d e l w h e r e s e lf is h u s e r s a r e c o m p e tin g f o r n e t-w o r k r e s o u r c e s u s in g e c o n o m ic a g e n ts to p u r c h a s e r e s o u r c e s in a n e tw o r k w h e r e n e tw o r k p r o v id e r s s e t p r ic e s b a s e d o n d e m a n d a n d a v a ila b le s u p p ly . • P a r i s M e t r o P r i c i n g ( P M P ) M o d e l P M P is a p a th - d if f e r e n tia te d s e r v ic e m o d e l b a s e d o n a u s a g e - s e n s itiv e p r ic in g s c h e m e p r e s e n te d b y O d ly z k o ( O d ly z k o , 1 9 9 9 ) . I t is b a s e d o n a " k e e p it a s s im p le a s it is " p r in -c ip le p r o p o s in g th e s e p a r a tio n o f th e I P b a c k b o n e in f r a s tr u c tu r e in to s e v e r a l lo g ic a lly s e p a r a te d c h a n n e ls , e a c h w ith a d if f e r e n t p r ic e p e r b y te a llo w in g u s e r a p p lic a tio n s to s e le c t f o r e a c h p a c k e t w h ic h c h a n n e l to s e n d it o n w ith th e e x p e c ta tio n th a t m o r e e x -p e n s iv e c h a n n e ls w o u ld a ttr a c t le s s tr a f f ic a n d th e r e f o r e w o u ld b e le s s c o n g e s te d . P M P is a s im p le a p p r o a c h a d d r e s s in g o n ly th e p r ic in g p a r t: n o a s s u m p tio n is m a d e a b o u t n e tw o r k o p tim iz a tio n in P M P e x c e p t th e n e e d to m a k e th e n e tw o r k a p p e a r a s s im p le a s p o s s ib le to th e u s e r s . • E x p e c t e d C a p a c i t y C o n t r a c t M o d e l T h e e x p e c te d c a p a c ity c o n tr a c t w a s p r o p o s e d b y C la r k ( C la r k , 1 9 9 6 ) a s a u s a g e - s e n s itiv e p r ic in g s c h e m e im p le m e n tin g a b a n d w id th a llo c a tio n m o d e l w h e r e u s e r s p a y a p r ic e f o r a h ig h p r o b a b ility o f d e liv e r y f o r a g iv e n v o lu m e o f tr a f f ic . I n th is a p p r o a c h u s e r a p p lic a tio n s a r e c h a r g e d f o r u s in g th e n e tw o r k c a p a c ity in s te a d o f b e in g c h a r g e d f o r a c tu a l u s a g e : u s e r a p p lic a tio n s w h o s e a c tu a l u s a g e e x c e e d s th e ir e x p e c te d c a p a c ity d u r in g c o n g e s tio n p e r io d s w ill e x p e r ie n c e a d e la y in th e ir tr a n s m is s io n in s te a d o f h a v in g to p a y m o r e th a n th e ir c o n tr a c te d c a p a c ity c o s ts .
3. Congestion-sensitive
pricing:
C o n g e s tio n p r ic in g is e m e r g in g a s a n e w p r ic in g a p p r o a c h e x te n s iv e ly in v e s tig a te d b y th e I n te r n e t c o m m u n ity c o m b in in g c o n g e s tio n a v o id a n c e w ith r e s o u r c e p r ic in g to a c h ie v e n e tw o r k o p tim a lity . C o n g e s tio n p r ic in g is b a s e d o n th e im p le -m e n ta tio n o f a n e q u a l- s e r v ic e f o r a ll u s e r s u n d e r lig h t a n d m o d e r a te lo a d c o n d itio n s a n d a d if f e r e n tia te d s e r v ic e b a s e d o n p r ic in g u n d e r h e a v y lo a d c o n d itio n s . • S m a r t M a r k e t P r i c i n g M o d e l S m a r t m a r k e t p r ic in g ( M a c k ie - M a s o n a n d V a r ia n , 1 9 9 5 ) p r o p o s e s a p a c k e t- le v e l p r ic in g m o d e l w h ic h im p le m e n ts a w illin g n e s s - to - p a y s c h e m e a llo w in g in d iv id u a l p a c k e ts to c a r r y th e p r ic e th e s e n d e r is w illin g to p a y to h a v e th a t p a c k e t s e n t s a f e ly . F r o m a s e t o fn
p a c k e ts tr a n s m itte d o n a lin k w h ic h c a n a c c o m m o d a te m p a c k e ts ( m ::;n),
a n e tw o r k im p le m e n tin g th e s m a r t m a r k e t w ill p r o c e s s th e m h ig h e s t b id p a c k e ts a n d d r o p th en -
m o th e r p a c k e ts . I n a s m a r t m a r k e t s c h e m e w h e r e th e m th h ig h e s t b id p a c k e ts a r e a d m itte d o n a lin k , th e m - 1 s t b id is id e n tif ie d to b e th e p r ic e to b e c h a r g e d to p a c k e ts 1 , ... , m . A s m a r t m a r k e t is in e q u ilib r iu m w h e n th e p r ic e p a id f o r s e n d in g a n a d d itio n a l p a c k e t r e f e r r e d to a s th e m a r g in a l c o s t is e q u a l to th e c u to f f p r ic e im p o s e d b y th e n e tw o r k . D e s p ite its r e la tiv e s im p lic ity , th e s m a r t m a r k e t s u f f e r s f r o m s e v e r a lim p le m e n ta tio n p ro b le m s in c lu d in g a h ig h tra n sa c tio n o v e rh e a d to im p le m e n t "m a rk e t c le a rin g " a t p o te n tia l b o ttle n e c k s in th e n e tw o rk , a p o ste rio ri p ric in g u n su ita b le to u se rs a n d th e p ro b le m o f p a c k e t re -o rd e rin g .
• Proportional
Fair Pricing
Model
P ro p o rtio n a l fa ir p ric in g (K e lly e t a I., 1 9 9 8 ) w a s p ro p o se d a s a v a ria n t o f c o n g e stio n c o n tro l m e c h a n ism s fo r th e T C P p ro to c o l. It b o rro w s th e m a rk e t-p ric in g c o n c e p t fro m th e sm a rt m a rk e t a n d th e a d d itiv e in c re a se /m u ltip lic a tiv e d e c re a se p ro p e rtie s fro m th e T C P p ro to c o l.
In
p ro p o rtio n a l fa ir p ric in g , th e n e tw o rk c h a rg e sp(y)
p e r u n it o f flo w tra n sm itte d a t a ra te o fy
p a c k e ts p e r u n it tim e .In
th is flo w -b a se d p ric in g a p p ro a c h , e a c h u se r se n d in g tra ffic in th e n e tw o rk a t a ra tex
is c h a rg e dxp(y).
T h a t u se r w ill a tte m p t to b rin g th e n e tw o rk 's c h a rg e in to e q u ilib riu m w ith its b id s b y a d ju stin g its tra n sm issio n ra te th ro u g h a n a d d itiv e in c re a se /m u ltip lic a tiv e d e c re a se fe e d b a c k c o n tro l sc h e m e .1.5
Thesis Outline and Main Contributions
V a rio u s fra m e w o rk s h a v e b e e n p ro p o se d b y th e th e re se a rc h c o m m u n ity to stu d y Q o S p ro v isio n in g m e c h a n ism s to b e d e p lo y e d in
IP
n e tw o rk s a t d iffe re n t tim e sc a le s. T h e tim e sc a le s in c lu d e lo n g -, m e d iu m - a n d sh o rt-te rm v a ria tio n s o f th e tra ffic . W e p ro p o se a n in te g ra te d a p p ro a c h c o m b in in g tra ffic re g u la tio n a n d tra ffic e n g in e e rin g m e c h a n ism s in a p a th -o rie n te d ro u tin g a rc h ite c tu re la y -e r-e d o v -e r a D iffs-e rv n -e tw o rk a s a so lu tio n to th e p ro b le m s o f tra ffic m a n a g e m e n t inIP
n e tw o rk s. T h is a p p ro a c h p ro p o se s tra ffic e n g in e e rin g fo r lo n g - a n d m e d iu m -te rm v a ria tio n s o f th e tra ffic a n d tra ffic re g u la tio n fo r sh o rt-te rm v a ria tio n s o f th e tra ffic .T h e e v o lu tio n o f
IP
Q o S a rc h ite c tu re s p re d ic ts a n e v o lu tio n o f th e In te rn e t in to a n e tw o rk in c lu d -in g tw o su b n e tw o rk s: a Q o S g u a ra n te e d su b n e tw o rk im p le m e n tin g c irc u it-sw itc h in g b a se d Q o S m e c h a n ism s u sin g a n In tse rv a rc h ite c tu re a n d a c o n tro lle d b e st-e ffo rt su b n e tw o rk la y e re d a b o v e a D iffse rv n e tw o rk w h e re p ric in g a n d a n d n o n -p ric in g m e c h a n ism s w ill b e d e p lo y e d to p ro v id e th e G o S re q u ire d b y d iffe re n t a p p lic a tio n s.T h e d e sig n o f a sc a la b le ro u tin g a rc h ite c tu re a n d th e lo c a liz a tio n o f d iffe re n t tra ffic m a n a g e m e n t m e c h a n ism s in th is ro u tin g e n v iro n m e n t is a se c o n d issu e a d d re sse d in th is th e sis. H o w tra ffic e n -g in e e rin -g a n d tra ffic re g u la tio n a re im p le m e n te d in th is a rc h ite c tu re is th e th ird issu e in v e stig a te d in th is th e sis.
1.5.1
Key Contributions
T h e m a in c o n trib u tio n s o f th is th e sis a re th re e fo ld .
• F irst, w e p ro p o se tra ffic a n d b a n d w id th a llo c a tio n m e c h a n ism s to b e d e p lo y e d fo r tra ffic e n g in e e rin g m u lti-se rv ic e lo g ic a l p a th s (M P L S L S P s). T h e se d iffe re n t te c h n iq u e s a re in -te g ra -te d in to a u n ifie d fra m e w o rk c o m b in in g lo a d -b a la n c in g a n d d iffe re n tia te d se rv ic e s to
1.5 Thesis Outline and Main Contributions
11
s u p p o rt n e tw o rk p ro v is io n in g a n d n e tw o rk re c o n fig u ra tio n to re s to re th e n e tw o rk o p tim a lity u n d e r p e rtu rb a tio n .
• T h e s e c o n d c o n trib u tio n c o n s is ts in th e d e s ig n o f a d e c e n tra liz e d ro u tin g a rc h ite c tu re th a t c o n tro ls th e n e tw o rk a t tw o le v e ls :
lo n g - a n d m e d iu m -te rm v a ria tio n s o f th e tra ffic u s in g g lo b a l c o n tro l m e c h a n is m s e x -e c u t-e d in a n e d g e -e d g e c o n tro l p la n e b a s e d o n a n e d g e -c o re in te ra c tio n to c o m p u te o p tim a l n e tw o rk p a ra m e te rs d u rin g n e tw o rk p ro v is io n in g (s e t-u p a n d re c o n fig u ra tio n ), a n d
s h o rt-te rm v a ria tio n s o f th e tra ffic u s in g lo c a l c o n tro l m e c h a n is m s d e p lo y e d in a u s e r-u s e r c o n tro l p la n e to re g u la te th e tra ffic e n te rin g th e n e tw o rk u s in g a u s e r-c o re in te r-a c tio n .
• T h e th ird c o n trib u tio n c o n s is ts in th e d e s ig n a n d e v a lu a tio n o f tra ffic re g u la tio n m e c h a n is m s d e p lo y e d in th e tw o c o n tro l p la n e s u s in g a n e a rly c o n g e s tio n d e te c tio n m o d e l b a s e d o n a th re s h o ld n e tw o rk lo a d in g s c h e m e to e ffe c t a fa ir a n d e ffe c tiv e a llo c a tio n o f th e n e tw o rk b a n d w id th to c o m p e tin g tra ffic flo w s .
1.5.2
Thesis
Outline
T h is th e s is is s u b d iv id e d in to tw o p a rts c o n ta in in g th e tw o m a in to p ic s o f
IP
tra ffic m a n a g e m e n t a d d re s s e d : T ra ffic e n g in e e rin g a n d tra ffic re g u la tio n . T ra ffic e n g in e e rin g a p p ro a c h e s a re p re s e n te d in th e firs t p a rt w h ic h in c lu d e s c h a p te r 2 a n d c h a p te r 3 . T h e s e c o n d p a rt in c lu d in g c h a p te r 4 a n d c h a p te r 5 c o n ta in s tra ffic re g u la tio n m e c h a n is m s .C h a p te r 2 a d d re s s e s tra ffic d is trib u tio n a s p e c ts o f tra ffic e n g in e e rin g . C h a p te r 3 p ro p o s e s a th re s h -o ld r-o u tin g a p p ro a c h to tra ffic e n g in e e rin g . T h e p ro p o s e d ro u tin g a rc h ite c tu re is p re s e n te d in c h a p te r 4 . C h a p te r 5 p re s e n ts m e c h a n is m s d e p lo y e d in th e re g u la tio n o f th e tra ffic u n d e r lo n g -, m e d iu m - a n d s h o rt-te rm v a ria tio n o f th e tra ffic p ro file . T h e c o n c lu s io n s a re p re s e n te d in c h a p te r 6 .
Chapter
2
Traffic
Engineering Model
2.1
Introduction
T r a f f i c e n g i n e e r i n g i s a n i m p o r t a n t a s p e c t o f n e t w o r k m a n a g e m e n t a l l o w i n g d a t a t o b e e f f i c i e n t l y r o u t e d t h r o u g h t h e n e t w o r k b y e f f e c t i n g Q o S a g r e e m e n t s b e t w e e n t h e a v a i l a b l e r e s o u r c e s a n d t h e c u r r e n t a n d e x p e c t e d t r a f f i c . M u l t i - p r o t o c o l l a b e l s w i t c h i n g ( M P L S ) e x t e n d s t h e I P v 4 d e s t i n a t i o n - b a s e d r o u t i n g p r o t o c o l s t o p r o v i d e n e w a n d s c a l a b l e r o u t i n g c a p a b i l i t i e s i n c l u d i n g t r a f f i c e n g i n e e r i n g a n d t r a f f i c f l o w d i f f e r e n t i a t i o n b a s e d o n e x p l i c i t r o u t i n g . M P L S w a s i n i t i a l l y p r o p o s e d b y t h e I E T F b a s e d o n a t o p o l o g y - d r i v e n r o u t i n g m o d e l w h i c h d e s p i t e i t s r e l a t i v e i m -p r o v e m e n t o v e r c o n v e n t i o n a l I P r o u t i n g b y p r o v i d i n g h i g h e r p r o c e s s i n g s p e e d s a n d d i f f e r e n t i a t i o n o f t r a f f i c f l o w s , i s c o n s t r a i n e d b y t h e s a m e s h o r t e s t - p a t h s e l e c t i o n t h a t a p p l i e s i n c o n v e n t i o n a l I P n e t w o r k s . C o n s t r a i n t - b a s e d r o u t i n g w a s n e x t p r o p o s e d a s a n i m p r o v e m e n t o v e r t h e b a s i c t o p o l o g y - d r i v e n M P L S t o s u p p o r t t r a f f i c e n g i n e e r i n g b y a l l o w i n g p a c k e t s b e l o n g i n g t o a f l o w t o b e s e n t t o a d e s t i n a t i o n u s i n g a r b i t r a r y p a t h s w h i c h w e r e s e l e c t e d b a s e d o n t h e f l o w 's c o n s t r a i n t s s u c h a s b a n d w i d t h g u a r a n t e e , l a t e n c y a n d j i t t e r m i n i m i z a t i o n o r o t h e r p e r f o r m a n c e r e q u i r e m e n t s . T h e s e v i r t u a l p a t h s a r e c a l l e d c o n s t r a i n t - r o u t e d L S P s ( C R - L S P s ) . C o n s t r a i n t - b a s e d r o u t i n g ( C R ) p r o v i d e s m o r e s c a l a b l e n e t w o r k o p e r a t i o n b y o f f - l o a d i n g t h e n e t w o r k a d m i n i s t r a t o r f r o m t h e t a s k o f m o n i t o r i n g t h e s t a t e o f t h e n e t w o r k a n d e x e c u t i n g r o u t i n g a n d c o m p e n s a t i o n m e c h a n i s m s w h e n p r o b l e m s a r i s e . C o n s t r a i n t - r o u t e d L S P s r a i s e t h r e e i m p o r t a n t i s s u e s : t h e i d e n t i f i c a t i o n o f t h e L S P s , t h e d i s t r i b u t i o n o f t h e t r a f f i c a m o n g t h e L S P s w h i c h p r e s e r v e s t h e Q o S r e q u i r e d f o r t h e t r a f f i c b e i n g r o u t e d t h r o u g h t h e n e t w o r k , a n d t h e s u p p o r t o f f a u l t - t o l e r a n c e a n d c o n g e s t i o n c o n t r o l . D i s t r i b u t i n g t h e n e t w o r k t r a f f i c a c r o s s n e t w o r k l i n k s w h i l e s u p p o r t i n g r e c o v e r y f r o m c o n g e s t i o n a n d f a u l t - t o l e r a n c e i s o n e o f t h e m a i n o b j e c t i v e s o f t r a f f i c e n g i n e e r i n g . T h i s c a n b e a c h i e v e d i n a m u l t i - p a t h e n v i r o n m e n t b y r e - r o u t i n g t h e t r a f f i c o v e r a l t e r n a t e p a t h s u p o n f a i l u r e o r c o n g e s t i o n . A l t h o u g h t h e r e - r o u t i n g a b i l i t y o f a n e t w o r k i n c r e a s e s w i t h t h e n u m b e r o f p a t h s a v a i l a b l e f o r r e - r o u t i n g , t h e i d e n t i f i c a t i o n o f a l a r g e n u m b e r o f p a t h s m a y b e p r o h i b i t i v e l y e x p e n s i v e . 1 3S u p p o r t i n g t r a f f i c e n g i n e e r i n g i n m u l t i - p a t h e n v i r o n m e n t s i s t h e r e f o r e a t a s k t h a t m u s t b e c a r r i e d o u t c a r e f u l l y t o a v o i d t h e p e r f o r m a n c e d e g r a d a t i o n t h a t c o u l d o c c u r f r o m a n i n e f f i c i e n t p a t h c o m p u t a t i o n a n d m a n a g e m e n t m e c h a n i s m . T h i s c h a p t e r a d d r e s s e s t h e f o l l o w i n g q u e s t i o n s r e l a t e d t o t h e o p t i m a l d i s t r i b u t i o n o f t r a f f i c f l o w s t o p r e - c o m p u t e d p a t h s i n a c o n s t r a i n t r o u t i n g e n v i r o n m e n t : • h o w i s a s e t o f p a t h s c o m p u t e d , • h o w i s a s u b s e t o f t h e c o m p u t e d p a t h s s e l e c t e d t o c a r r y t h e o f f e r e d t r a f f i c , • h o w i s t h e o f f e r e d t r a f f i c d i s t r i b u t e d a m o n g t h e s e l e c t e d p a t h s , a n d • h o w u s e f u l a r e t h e r e m a i n i n g u n s e l e c t e d p a t h s a s b a c k u p s t o b e u s e d i n t h e e v e n t o f t r a f f i c o v e r l o a d a n d / o r p a t h f a i l u r e W e p r e s e n t t h r e e p a t h i d e n t i f i c a t i o n s c h e m e s t o f i n d p a t h s e t s . A s u b s e t o f t h e s e p a t h s - t h e b e s t p a t h s i n t e r m s o f t h e l i n k c o s t m e t r i c s - i s u s e d a s a c t i v e p a t h s t o c a r r y t h e o f f e r e d t r a f f i c . T h e u n s e l e c t e d p a t h s a r e r e s e r v e d a s b a c k - u p p a t h s t o c a r r y r e - r o u t e d t r a f f i c . W e d e s c r i b e t w o t r a f f i c d i s t r i b u t i o n m e t h o d s f o r l o a d - b a l a n c i n g t h e s e t o f a c t i v e p a t h s f o r c a r r y i n g t h e t r a f f i c o f f e r e d t o a n e t w o r k a n d a s e r v i c e d i f f e r e n t i a t i o n m o d e l b a s e d o n t r a f f i c p r i o r i t i z a t i o n . T h e r e m a i n d e r o f t h i s c h a p t e r i s o r g a n i z e d a s f o l l o w s .
In
s e c t i o n 2 . 2 w e d e s c r i b e t h e b a s i c f l o w o p -t i m i z a -t i o n m o d e l . s e c t i o n 2 . 3 d e s c r i b e s s o l u t i o n c h a r a c t e r i s t i c s . S e c t i o n 2 . 4 p r e s e n t s e x p e r i m e n t a l r e s u l t s . T h e c o n c l u s i o n s a r e p r e s e n t e d i n s e c t i o n 2 . 5 .2.2
A Basic Flow Optimization Model
P a t h i d e n t i f i c a t i o n a n d t r a f f i c d i s t r i b u t i o n a r e t h e t w o m o s t i m p o r t a n t p r o c e s s e s i n v o l v e d i n f l o w o p t i m i z a t i o n . A f l o w o p t i m i z a t i o n m o d e l m a y b e b a s e d o n e i t h e r a r e a c t i v e s c h e m e w h e r e p a t h i d e n t i f i c a t i o n a n d t r a f f i c d i s t r i b u t i o n a r e c o m p u t e d s i m u l t a n e o u s l y t o a c h i e v e a n o p t i m a l t r a f f i c f l o w a l l o c a t i o n , o r a p r e - p l a n n e d m o d e l w h e r e p a t h i d e n t i f i c a t i o n a n d t r a f f i c d i s t r i b u t i o n a r e p e r f o r m e d s e p a r a t e l y . W e c o n s i d e r a f l o w o p t i m i z a t i o n m o d e l b a s e d o n t h r e e c o n c e p t s : t h e s e p a r a t i o n b e t w e e n p a t h i d e n t i f i c a t i o n a n d t r a f f i c d i s t r i b u t i o n , t h e i n c r e m e n t a l a s s i g n m e n t o f t h e o f f e r e d t r a f f i c t o i d e n t i f i e d p a t h s , a n d t h e s e p a r a t i o n o f t h e s e t o f i d e n t i f i e d p a t h s i n t o a c t i v e p a t h s a n d b a c k - u p p a t h s . W e p r e s e n t a p r e - p l a n n e d f l o w o p t i m i z a t i o n m o d e l w h i c h c o n s i s t s o f t h r e e s t e p s ;
Path identification.
P a t h s e t s a r e c o m p u t e d f o r e a c h I - E ( I n g r e s s - E g r e s s ) p a i r u s i n g o n e o f t h r e e m e t h o d s .2 .2 A B a s ic F lo w O p tim iz a tio n M o d e l
15
Path selection
and activation.
T h e p a th s e t is p a r titio n e d to y ie ld a s e t o f a c tiv e p a th s to c a r r y th e o f f e r e d tr a f f ic a n d a s e t o f b a c k - u p p a th s f o r f a u lt r e c o v e r y a n d /o r c o n g e s tio n a v o id a n c e . A c tiv e p a th s a r e id e n tif ie d e ith e r s ta tic a lly b y s e le c tin g a f ix e d s e t o f o f a c tiv e p a th s o r d y n a m ic a lly b y p r e - s e le c tin g a f e w a c tiv e p a th s a n d a c tiv a tin g n e w p a th s w h e n th e o f f e r e d tr a f f ic c a n n o t b e a c c o m m o d a te d b y th e a v a ila b le a c tiv e p a th s .Traffic distribution.
W e c o n s id e r tw o m o d e ls : a p r o p o r tio n a l tr a f f ic d is tr ib u tio n m o d e l w h ic h s h a r e s th e o f f e r e d tr a f f ic a m o n g th e p a th s a c c o r d in g to w e ig h ts e x p r e s s in g th e lo a d in g o f th e s e p a th s , a n d a f lo w d e v ia tio n m o d e l ~ a v a r ia n t o f th e s ta n d a r d f lo w d e v ia tio n a lg o r ith m - w h ic h m o v e s th e o f f e r e d tr a f f ic f r o m th e h ig h e s t c o s t p a th s to lo w e r c o s t p a th s .2 .2 .1
P a th
I d e n tif ic a tio n
W e p r e s e n t tw o m e th o d s f o r f in d in g p a th s : a
K
s h o r te s t p a th ( K S P ) m e th o d a n d a to p o lo g y - b a s e d s h o r te s t p a th ( T S P ) m e th o d .A p a th
p
c o n n e c tin g a n I - E p a ir (i,e ) is a n o n - c y c lic s e q u e n c e o f lin k s (£ 1 , £ 2 , ... ,£h)
c o n n e c tin g th e in g r e s s n o d ei
to th e e g r e s s n o d e e . F o r e a c h I - E p a ir , th e K S P m e th o d f in d s u p toK
lin k - d is jo in t s h o r te s t p a th s p a s s in g th r o u g h th eK
n o d e s a d ja c e n t to th e in g r e s s n o d e . S in c e th e s e p a th s a r e lin k - d is jo in t, th e K S P m e th o d p r o d u c e s d if f e r e n t p a th s e ts d e p e n d in g u p o n th e o r d e r u s e d f o r th e s u c c e s s iv eK
p a th d is c o v e r y . C o m p u tin g th e lin k - d is jo in t p a th s p a s s in g th r o u g h th e in g r e s s n e ig h b o u r s s e le c te d in a s c e n d in g , d e s c e n d in g o r r a n d o m le x ic o g r a p h ic o r d e r c a n p r o d u c e th r e e d if f e r e n t p a th s e ts . T h is is illu s tr a te d b y th e " tr a p " n e tw o r k p r e s e n te d in F ig u r e 2 .1 . T h o u g h th e r e a r e th r e e p a th s f r o m n o d e 1 to n o d e 4 , th e K S P a lg o r ith m w ill f in d o n ly o n e p a th if p a th r(1,2,3,4) is s e le c te d f ir s t. O th e r w is e , a t m o s t tw o p a th s w ill b e f o u n d b y th e K S P a lg o r ith m . F ig u r e 2 .1 : T h e " tr a p " n e tw o r k T h e T S P m e th o d f in d s th eK
s h o r te s t p a th s p a s s in g th r o u g h th eK
n o d e s a d ja c e n t to th e in g r e s s n o d e . D e s p ite its r e la tiv e s im p lic ity , th is m e th o d m a y p r o d u c e p o o r p a th s e ts s in c e d if f e r e n t p a th s m a y s h a r e lin k s . N o n - lin k - d is jo in t p a th s - m a y p r o d u c e p o o r p e r f o r m a n c e in f a u lt r e c o v e r y a n d p r e s e n t tr a f f ic b o ttle - n e c k s u n d e r h e a v y lo a d s .2.2.2
Path Selection and Activation
Active Paths
P a th id e n tif ic a tio n m a y p r o d u c e m a n y p a th s b e tw e e n e a c h 1 -E p a ir . H o w e v e r , p r a c tic a l r o u tin g e n v ir o n m e n ts u s u a lly u s e o n ly a s u b s e t o f p a th s to c a r r y th e o f f e r e d tr a f f ic . T h e p a th s a r e th e r e f o r e r a n k e d a c c o r d in g to s e le c te d m e tr ic s in o r d e r to d if f e r e n tia te b e tw e e n a c tiv e p a th s a n d b a c k -u p p a th s . T h e m e tr ic s a r e : lin k c a p a c ity ; h o p c o u n t - th is m e tr ic a llo w s th e s e le c tio n o f le s s r e s o u r c e c o n s u m in g p a th s r e p r e s e n te d b y s h o r te r p a th s ; lin k in te r f e r e n c e e x p r e s s e d b y th e n u m b e r o f p a th s tr a v e r s in g th is lin k - th is m e tr ic a llo w s a r a n k in g o f p a th s w h ic h r e d u c e s th e p r o b a b ility o f p r o d u c in g b o ttle n e c k p a th s in th e n e tw o r k ; a n d o th e r m e tr ic s e x p r e s s in g r o u tin g p o lic ie s o r a c o s t f u n c tio n c o m b in in g tw o o r m o r e c o s t f u n c tio n s . W e c la s s if y th e p a th s a c c o r d in g to th e ir le n g th a n d e f f e c tiv e b a n d w id th u s in g th e f o llo w in g p r e f -e r -e n c -e f u n c tio n s G1
(p)
G
2(p)
l/L~
Cp/L~
(1)
w h e r e
C
p is th e e f f e c tiv e b a n d w id th o f p a thp
e x p r e s s e d b y th e m in im u m o v e r its lin k 's c a p a c itie s ,L
p is th e le n g th o f th e p a th e x p r e s s e d b y th e n u m b e r o f h o p s in p a thp
a n d 0 :: is a p o s itiv e v a lu ee x p r e s s in g a p e n a lty r e la te d to th e le n g th o f p a th
p.
T h e p r e f e r e n c e f u n c tio n G1 is b e tte r s u ite d to tr a f f ic d is tr ib u tio n s c h e m e s w h e r e th e tr a f f ic is d is
-tr ib u te d a m o n g th e le a s t le n g th p a th s . T h e p r e f e r e n c e f u n c tio n G2 is b e tte r s u ite d to tr a f f ic d is tr i-b u tio n s c h e m e s w h e r e th e tr a f f ic is d is tr ib u te d a m o n g th e le a s t le n g th p a th s a n d o v e r - p r o v is io n e d p a th s .
Back-up Paths
T h e p r e f e r e n c e f u n c tio n s ( 1 ) a r e u s e d to p a r titio n th e s e t o f p a th s in to tw o p a th s e ts c o r r e s p o n d in g to tw o lo g ic a l n e tw o r k s : a n a c tiv e n e tw o r k c o m p o s e d o f th e s e t o f n o d e s a n d lin k s tr a v e r s e d b y th e p a th s th a t h a v e b e e n a llo c a te d tr a f f ic - th e s e p a th s a r e c a lle d a c tiv e p a th s , a n d a b a c k - u p n e tw o r k c o m p o s e d o f n o d e s a n d lin k s tr a v e r s e d b y p a th s th a t h a v e b e e n r e s e r v e d f o r f a ilu r e r e c o v e r y a n d c o n g e s tio n c o n tr o l.
B a c k - u p p a th s m a y b e u s e d e ith e r a s e m e r g e n c y p a th s o r r e p la c e m e n t p a th s w h e n th e n e tw o r k e x p e r ie n c e s a f a ilu r e o r c o n g e s tio n . T h e s e p a th s a r e a c tiv a te d w h e n o n e o r m o r e a c tiv e p a th s e x p e r ie n c in g a lin k o r n o d e f a ilu r e h a v e to b e r e p la c e d , a n d /o r th e n u m b e r o f p a th s c a r r y in g th e tr a f f ic f o r a g iv e n I - E p a ir a r e n o t s u f f ic ie n t to c a r r y th e o f f e r e d tr a f f ic .