Optimizing Cost and Performance for Multihoming

1
Optimizing Cost and Performance for Multihoming
Lili QiuMicrosoft Research liliq_at_microsoft.com
Joint Work withD. K. Goldenberg, H. Xie, Y. R.
Yang, Yale University Y. Zhang, ATT Labs
Research
ACM SIGCOMM 2004
2
Multihoming Smart Routing

• Multihoming
• A popular way of connecting to Internet
• Smart routing
• Intelligently distribute traffic among multiple
  external links

3
Potential Benefits

• Improve performance
• Potential improvement 25 Akella03
• Similar to overlay routing Akella04
• Improve reliability
• Two orders of magnitude improvement in fault
  tolerance of end-to-end paths Akella04
• Reduce cost

Q How to realize the potential benefits?
4
Our Goals

• Goal
• Design effective smart routing algorithms to
  realize the potential benefits of multihoming
• Questions
• How to assign traffic to multiple ISPs to
  optimize cost?
• How to assign traffic to multiple ISPs to
  optimize both cost and performance?
• What are the global effects of smart routing?

5
Related Work

• Techniques for implementing multihoming
• BGP peering, DNS-based, NAT-based (e.g.,
  RFC2260, Cisco, GCLC04, Radware, F5)
• Complementary to our work
• Performance evaluation Akella03,Akella04
• Quantify the potential benefits of multihoming
• Unaddressed challenge how to achieve this in
  practice
• Smart routing
• Commercial products (e.g., RouteScience,
  Internap, Proficient, )
• Technical details are unavailable
• Hash-based load balancing Cao01, Guo04
• Optimizes neither performance nor cost

6
Network Model

• Network performance metric
• Latency (also an indicator for reliability)
• Extend to alternative metrics
• log (1/(1-lossRate)), or latencywlog(1/(1-lossRa
  te))
• ISP charging models
• Cost C0 C(x)
• C0 a fixed subscription cost
• C a piece-wise linear non-decreasing function
  mapping x to cost
• x charging volume
• Total volume based charging
• Percentile-based charging (95-th percentile)

7
Percentile Based Charging
Sorted volume
Interval
N
95N
Charging volume traffic in the (95N)-th sorted
interval
8
Why cost optimization?

• A simple example
• A user subscribes to 4 ISPs, whose latency is
  uniformly distributed
• In every interval, the user generates one unit of
  traffic
• To optimize performance
• ISP 1 1, 0, 0, 0,
• ISP 2 0, 1, 0, 0,
• ISP 3 0, 0, 1, 0,
• ISP 4 0, 0, 0, 1,
• 95th-percentile 1 for all 4 ISPs
• 95th-percentile 1 using one ISP
• Cost(4 ISPs) 4 cost(1 ISP)

Optimizing performance alone could result in high
cost!
9
Cost Optimization Problem Specification (2 ISPs)
Volume
Time
N
1
2
10
Cost Optimization Problem Specification (2 ISPs)
Sorted volume
Volume
P1
Sorted volume
Time
P2
Goal minimize total cost C1(P1)C2(P2)
11
Issues Insights

• Challenge traditional optimization techniques do
  not work with percentiles
• Key determine each ISPs charging volume
• Results
• Let V0 denote the sum of all ISPs charging
  volume
• Theorem 1 Minimize cost ?? Minimize V0
• Theorem 2 V0 1- ?k1..N(1-qk) quantile of
  original traffic, where qk is ISP ks charging
  percentile

12
Cost Optimization Problem Specification (2 ISPs)
Sorted volume
Volume
P1
Sorted volume
Time
P2
P1 P2 ? 90-th percentile of original traffic
13
Intuition for 2-ISP Case

• ISP 1 has ? 5 intervals whose traffic exceeds P1
• ISP 2 has ? 5 intervals whose traffic exceeds
  P2
• The original traffic (ISP 1 ISP 2 traffic) has
  ? 10 intervals whose traffic exceeds P1P2
• P1P2 ? 90-th percentile of original traffic

14
Sketch of Our Algorithm

• Determine charging volume for each ISP
• Compute V0
• Find pk that minimize ?k ck(pk) subject to
  ?kpkV0 using dynamic programming
• Assign traffic given charging volumes
• Non-peak assignment ISP k is assigned ? pk
• Peak assignment
• First let every ISP k serve its charging volume
  pk
• Dump all the remaining traffic to an ISP k that
  has bursted for fewer than (1-qk)N intervals

15
Additional Issues

• Deal with capacity constraints
• Perform integral assignment
• Similar to bin packing (greedy heuristic)
• Make it online
• Traffic prediction
• Exponential weighted moving average (EWMA)
• Accommodate prediction errors
• Update V0 conservatively
• Add margins when computing charging volumes

16
Optimizing Cost Performance

• One possible approach design a metric that is a
  weighted sum of cost and performance
• How to determine relative weights?
• Our approach optimize performance under cost
  constraints
• Use cost optimization to derive upper bounds of
  traffic that can be assigned to each ISP
• Assign traffic to optimize performance subject to
  the upper bounds

17
Evaluation Methodology

• Traffic traces (Oct. 2003 Jan. 2004)
• Abilene traces (NetFlow data on Internet2)
• RedHat, NASA/GSFC, NOAA Silver Springs Lab, NSF,
  National Library of Medicine
• Univ. of Wisconsin, Univ. of Oregon, UCLA, MIT
• MSNBC Web access logs
• Realistic cost functions Feb. 2002 Blind RFP
• Delay traces
• NLANR traces 3 months RTT measurements between
  pairs of 140 universities
• Map delay traces to hosts in traffic traces

18
Baseline Algorithms

• Round robin
• In each interval, assign traffic to a single ISP
• Rotate in a round robin fashion
• Equal split
• In each interval, split traffic equally among
  ISPs
• Similar to hash-based load balancing
• Offline local fractional
• Minimize the total cost for each interval
  independently
• Dedicated links
• Flat rate and independent of usage

19
Cost Comparison for Different Traces
Our algorithms significantly out-perform the
alternatives.
20
Cost Comparison for Varying Links
For all ISPs, our cost optimization performs
well.
21
Cost Performance Evaluation
Optimizing performance alone often doubles the
cost.
22
Cost Performance Evaluation (Cont.)
Our dual metric optimization achieves low cost
and latency.
23
Global Effects of Smart Routing

• Selfish nature of smart routing
• Each user optimizes its own cost performance
  without considering its impact on other traffic
• Need to understand its global effects
• Questions
• How well does smart routing perform when traffic
  assignment affects link latency?
• How well do different smart routing users
  co-exist?
• How well do smart routing users co-exist with
  single-homed users?

24
Evaluation Methodology

• Abilene traffic traces
• Rocketfuel inter-domain topology
• 170 nodes, 600 edges
• With propagation delay and OSPF weights
• M/M/1 queuing model
• Routing
• A user selects best performing ISP subject to
  cost constraints
• Inter-domain shortest AS hop count
• Intra-domain OSPF
• Compute traffic equilibria as in QYZS03

25
Global Effects Summary

• Impact of self interference is small
• Smart routing users co-exist well with each other
• Smart routing users co-exist well with
  single-homed users

26
Conclusions

• Contributions
• First paper on jointly optimizing cost and
  performance for multihoming
• Propose a series of novel smart routing
  algorithms that achieve both low cost and good
  performance
• Under traffic equilibria, smart routing improves
  performance without hurting other traffic
• Future work
• Further evaluation through Internet experiments
• Dynamics of interactions among different users
• Design better charging models

27

• Thank you!