Can Congestion Control and Traffic Engineering be at Odds?

1
Can Congestion Control and Traffic Engineering be
at Odds?

• Jiayue He, Mung Chiang, Jennifer Rexford
• Princeton University
• November 30th, 2006

2
Motivation

• Congestion Control
• maximize user utility

• Traffic Engineering
• minimize network congestion

Given routing Rli how to adapt end rate xi?
Given traffic xi how to perform routing Rli?
3
Congestion Control Model
Users are indexed by i
aggregate utility
Utility Ui(xi)
max. ? i Ui(xi) s.t. ?i Rlixi cl var. x
capacity constraints
Source rate xi
Congestion control provides fair rate allocation
amongst users
KellyMaullooTan98, Low03, Srikant04
4
Traffic Engineering Model
Links are indexed by l
aggregate cost
Cost f(yl/cl)
yl cl
min. ?l f(yl/cl) s.t. yl ?i Rlixi var. R
Link Load yl
Traffic engineering avoids bottlenecks in the
network
FortzThorup02, Rexford06
5
Model of Internet Reality
Congestion Control max ?i Ui(xi), s.t. ?i Rlixi
cl
xi
Rli
Traffic Engineering min ?l f(yl/cl), s.t. yl
?i Rlixi
6
System Properties

• Convergence
• Does it achieve some objective?
• Benchmark
• Utility gap between the joint system and benchmark

max. ?i Ui(xi) s.t. Rx c Var. x, R
WangLiLowDoyle05, HeChiangRexford06
7
Numerical Experiments

• System converges
• Quantify the gap to optimal aggregate utility
• Capacity distribution truncated Gaussian with
  average 100
• 500 points per standard deviation

Access-Core
Abilene Internet2
8
Results for Access-Core
Homogeneous optimal
Aggregate utility gap
Utility gap can exist
Standard deviation
Homogenous capacity reduces gap
9
Results for Abilene
Aggregate utility gap
Gap exists
Standard deviation
10
Abilene Continued f n(yl/cl)n
Aggregate utility gap
n
Gap shrinks with larger n
11
Backward Compatible Design

• Simulation of the joint system suggests that it
  is stable, but suboptimal
• Gap reduced if we modify f

f(yl/cl)
Cost f
f(yl/cl)
yl cl
0
Link load yl
12
Theoretical Results

• Modify congestion control to approximate the
  capacity constraint with a penalty function
• Theorem modified joint system model converges if
  Ui(xi) -Ui(xi) /xi

Master Problem min. g(x,R) - ?iUi(xi)
??lf(yl/cl)
Congestion Control argminx g(x,R)
Traffic Engineering argminR g(x,R)
13
Caveat

• Changing f allows for maximizing aggregate user
  utility
• Bottleneck links created
• Fragile to high volume traffic bursts
• Robustness lost

14
Conclusions So Far

• Model interaction between congestion control and
  traffic engineering
• Confirm intuition of the operators
• Stable
• Robust
• Modified joint system
• Optimal but not robust

15
New Objective
Congestion Control User Performance
Traffic Engineering Network Robustness
Can be at odds!

• To balance performance and robustness
• New objective max. ?iUi(xi) - ?lf(yl/cl)

16
Ongoing work

• DATE online distributed solution to new
  objective
• J. He, M. Bresler, M. Chiang and J. Rexford.
  Towards Robust Multilayer Traffic Engineering"
  In submission to JSAC Special Issue on
  Cross-layer Traffic Engineering.
• www.princeton.edu/jhe/

Congestion price

• Links
• Update prices
• Update effective capacity

• Edge router
• Rate limits incoming traffic
• Performs multipath routing

Link load
17
Future work

• Prove stability of joint system by modeling as a
  two player game
• Consider topology changes
• Link failures
• Mobile nodes
• Multi-domain version

18
The End

• Thank you!