Routing Bottlenecks in the Internet: Causes, Exploits, and - - PowerPoint PPT Presentation

Routing Bottlenecks in the Internet: Causes, Exploits, and Countermeasures

Min Suk Kang Virgil D. Gligor

ECE Department and CyLab, Carnegie Mellon University

Nov 4, 2014

2

Route Diversity is Critical to Resiliency of Internet Connectivity

rest of the world geographic area with poor route diversity

… …

link-flooding attack

3

Fortunately, most countries have enough route diversity

(source: www.renesys.com/2014/02/internetunderfire/)

40 + ≤ 40 ≤ 10 ≤ 2 # of ISPs with direct international connectivity

Then, do we need to worry about the link-flooding attacks?

Most countries have 10+ ISPs with international connections => good Internet route diversity

Unfortunately, YES.

• Why?

4

Despite high route diversity, Internet connectivity of countries can be degraded

Paper illustrates 1. pervasive phenomenon of routing bottlenecks 2. causes of routing bottlenecks 3. impact of targeted attacks & countermeasures the vast majority of Internet routes to chosen destinations concentrated on a small set of links routing bottleneck

5

sources (S)

mincut, M(S,D) routing bottleneck, B |B| ≪ |M(S,D)| e.g. 10 ≪ 1000

Mincut and Routing Bottleneck

routing bottleneck ≠ bandwidth bottleneck

geographic area destinations (D)

traceroute

6

M(S,D) B

Routing Bottlenecks in the current Internet

Normalized Link Occurrence  high rank low rank  Rank of Links in M(S,D)

0.02 0.04 0.06 0.08 0.1 0.12 200 400 600 800 1000 1200 0.02 0.04 0.06 0.08 0.1 0.12 10 20 30 40 50

B

1,000 randomly selected working servers

sources (S)

250 nodes in PlanetLab

(in 164 cities in 39 countries) geographic area

measurement for a country

destinations (D)

Link Occurrence (ratio)

(0.80)

0.000001 0.00001 0.0001 0.001 0.01 0.1 1 1 10 100 1000 10000

7

Routing Bottlenecks in 15 Countries

0.000001 0.00001 0.0001 0.001 0.01 0.1 1 1 10 100 1000 10000 0.000001 0.00001 0.0001 0.001 0.01 0.1 1 1 10 100 1000 10000 Country1 Country2 Country3 Country4 Country5 Country6 Country7 Country8 Country9 Country10 Country11 Country12 Country13 Country14 Country15 Country1 Country2 Country3

Normalized link occurrence Rank of Link

• link occurrence is accurately modeled by a power-law

Country1 Country15

α = 1.31 (β = 7.8) α = 2.36

Tested Countries (alphabetical) Brazil Egypt France Germany India Iran Israel Italy Japan Romania Russia

• S. Korea

Taiwan Turkey UK

Zipf-Mandelbrot distribution

f(k) = 1 (k + β)α

8

Routing Bottlenecks in 15 Large Cities

0.000001 0.00001 0.0001 0.001 0.01 0.1 1 1 10 100 1000 10000 0.000001 0.00001 0.0001 0.001 0.01 0.1 1 1 10 100 1000 10000 0.000001 0.00001 0.0001 0.001 0.01 0.1 1 1 10 100 1000 10000 City1 City2 City3 City4 City5 City6 City7 City8 City9 City10 City11 City12 City13 City14 City15 City1 City2 City3

Normalized link occurrence Rank of Link City1 City15

α = 1.38 (β = 7.8) α =2.17

• link occurrence is accurately modeled by a power-law

Tested Cities (alphabetical) Beijing Berlin Chicago Guangzhou Houston London Los Angeles Moscow New York Paris Philadelphia Rome Shanghai Shenzhen Tianjin

Zipf-Mandelbrot distribution

f(k) = 1 (k + β)α

routers Internet

…

route construction: sentence construction:

Speaker

9

Causes?

“Principle of least effort” [Zipf’49, Mandelbrot’53] conjecture: route-cost minimization policies

link1 link2 linkn … word1 word2 wordn …

==> Z-M distribution of word occurrence ==> Z-M distribution of link occurrence

An Analogy w/ Word Occurrence Distribution…

Policy: route-cost minimization

– BGP favors minimum-cost link => AS-level route concentration

10

Evidence for Inter-Domain Routing

AS*

(*) AS: autonomous system

AS AS AS $$$ $

• Test:

– policy I: favors min-cost links – policy II: distribute routes uniformly

AS

Rank of Inter-AS Links

• Norm. Link Occurrence

?

• Practice: route-cost minimization

– hierarchical topology + shortest path routing => route concentration at backbones

11

• Test: all possible ingress/egress

routes

– clear Zipf-Mandelbrot distribution

AS

Rank of Intra-AS Links

• Norm. Link Occurrence

Evidence for Intra-Domain Routing

3 link locations:

AS2 IXP AS1 … … … … … AS3 AS4

(Internet exchange points)

12

Link Types of Routing Bottlenecks

intra-AS inter-AS IXP

3 AS categories: Tier-1

(Global Transits/ National Backbones) (regional providers) (customers)

Tier-2 Tier-2

Tier-3 Tier-3 Tier-3 Tier-3

13

Link Types of Routing Bottlenecks

• various link types: intra (30%), inter (30%), and IXP (20%)
• 91% of inter/intra-AS links are owned by Tier-1/Tier-2

< Avg. link types of 50 bottleneck links of 15 countries (percentage) >

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

IXP Not Deter. INTRA-AS INTER-AS

in in

Tier-1 Tier-2

in in in inter ( inter ( inter (

Tier1-Tier1 Tier1-Tier2 Tier1-Tier3 Tier2-Tier2 Tier2-Tier3

Not Determined

14

Routing-Bottleneck Exploits Massive Link Flooding

e.g., Crossfire attack [IEEE S&P 2013] Link-flooding with indistinguishable attack flows Botnets Decoy Servers

… …

low-rate attack flows (e.g., 40 Gbps = 4 Kbps x 10K bots x 1K decoys) routing-bottleneck link Target Geographic area

several hops away

15

Connectivity Degradation in 15 Countries

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 10 20 30 40 50 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 10 20 30 40 50 Country1 Country2 Country3 Country4 Country5 Country6 Country7 Country8 Country9 Country10 Country11 Country12 Country13 Country14 Country15

Country1 Country15

Number of Links to Flood Degradation Ratio

α = 2.36 (β = 7.8) α = 1.31

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 10 20 30 40 50

16

Connectivity Degradation in 15 Large Cities

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 10 20 30 40 50 City1 City2 City3 City4 City5 City6 City7 City8 City9 City10 City11 City12 City13 City14 City15

City1 City15

Number of Links to Flood Degradation Ratio

α = 2.17 (β = 7.8) α = 1.38

17

Countermeasures

AS1 AS2 AS1 AS2 AS3 AS4

• Inter-domain links
• Load balancing across links to

different ASes [SIGCOMM’06]

• Load balancing across parallel

links between two ASes [ATC’07]

18

Countermeasures

AS

• Intra-domain links
• MPLS tunnels
• Equal-cost multipath (ECMP)

1 1 2

AS

Needs real-time link-weight re-adjustment Needs real-time MPLS traffic enginnering

(unknown if recent SDN-based solutions can be applied here)

19

Effectiveness of Countermeasures

10 20 30 40 50 60 70 80 90 100

Reduction of degradation ratio (%)

10 20 30 40 50 60 70 80 90 100

Reduction of degradation ratio (%)

10 20 30 40 50 60 70 80 90 100

Reduction of degradation ratio (%)

10 20 30 40 50 60 70 80 90 100

Reduction of degradation ratio (%)

4 implementation alternatives:

• “one type fits all” countermeasures are not very effective
• countermeasures at large ISPs (Tier-1&2) are most effective

Inter-AS links Intra-AS links Tier-1 ASes Tier-1&2 ASes

20

Related Work

• Internet topology studies; e.g., CAIDA, DIMES, etc.
• Power-law in Internet connectivity; e.g., [SIGCOMM’99, NATURE’00]
• Link-flooding attacks; e.g., Coremelt [ESORICS’09], Crossfire [S&P’13]

• Notion of the routing bottlenecks

– they are pervasive (in 15 countries and 15 cities)

21

Conclusions

• Causes: route-cost minimization

– very desirable feature of Internet routing

• Countermeasures

– effective when implemented in large ISPs

22

Thank You

Min Suk Kang (minsukkang@cmu.edu)