abilene resiliency test

Abilene Resiliency Test

John MooreTechnical Director

Centaur Lab

North Carolina ITEC

jhm@nc-itec.org

Acknowledgements

• SDSC– Kevin Walsh– Tom Hutton

• Abilene NOC– Andrew Lee– Matt Davy

• CENIC– Erick Sizelove

• Spirent Communications

• Qwest• Internet2

– Jeff Boote– Chris Heermann

Ominous?

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

Motivations

• OC-192 backbone links from Qwest “unprotected”– Previous OC-48 backbone circuits had

SONET protection (sub- 50 ms)

• Ongoing MPLS investigation requires understanding of various protection options available

Test Objectives• Measure convergence time after a link outage

– End-to-end and backbone edge-to-edge– Use longest convenient path

• SDSC to NCSU

• Examine effect of modifying IGP settings– Chose spf delay value as variable

• Examine different MPLS recovery schemes– Soft LSP (loose hop)– Explicit route (strict hops)– Fast re-route

Test Traffic• Spirent Communications AX/4000 tester in

CA-ITEC and NC-ITEC Labs– One IPv4 stream– Five MPLS streams– Unidirectional

• AMI owamp streams between LOSA and WASH– Actually, we have data for any-any Abilene nodes

• IS-IS monitor tool– Not deployed in time

Test Traffic• Owamp traffic

– 64 bytes, 10 pps• Increased rate for test, will leave as new default

– Measure latency and packet loss (indirectly)

• AX/4000 traffic– 1000 bytes frames, 20 Mbps per stream– MPLS tunnel traffic was Ethernet

• Used Juniper L2 circuits to get vlan-tagged traffic into LSP

– Measure packet rate and packet loss

IS-IS Parameters• Link failure hold time

– Delay in declaring link down– Default: zero

• LSP interval– Rate limit for LSP flooding (per interface)– Default: 30 LSPs per second

• SPF delay– Delay in running spf algorithm – Default: 200 ms

• SPF recalculation time– Time required to run spf algorithm– Abilene: usually < 1 ms, some spikes to 2 ms

SNVA

LOSA

STTL

CA-ITEC NC-ITEC

HSTN

DNVR KSCY IPLS

ATLA

CHIN NYCM

WASH

Primary BB Path

Secondary BB Path

NCNICalREN

Primary & Backup BB Paths

IP traffic - spf delay = 200 ms

0

500

1000

1500

2000

2500

3000

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9

Seconds

Packet Rate (pps) Packet Loss (packets)

IP traffic - spf delay = 50 ms

0

500

1000

1500

2000

2500

3000

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9

Seconds

Packet Rate (pps) Packet Loss (packets)

WASH-LOSA - spf delay = 200 ms

32

34

36

38

40

42

44

46

48

50

52

6.5 6.7 6.9 7.1 7.3 7.5 7.7 7.9 8.1 8.3 8.5 8.7 8.9 9.1 9.3 9.5 9.7 9.9 10.1 10.3 10.5

Time (seconds)

Late

ncy (

millisecon

ds)

SNVA

LOSA

STTL

CA-ITEC NC-ITEC

HSTN

DNVR KSCY IPLS

ATLA

CHIN NYCM

WASH

NCNICalREN

Stitched Path

End-to-end Path

MPLS Path Structure

MPLS Paths

Path Type Backup type Backup path

A End-end Soft LSP Via IGP

B End-end Explicit route Explicit

D Stitched Soft LSP Via IGP

E Stitched Explicit route Explicit

F Stitched Fast reroute Pre-computed across failed link

Path B - spf delay = 200 ms

0

500

1000

1500

2000

2500

3000

3500

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9

Seconds

Packet Rate (pps) Packet Loss (packets)

Path B - spf delay = 50 ms

0

500

1000

1500

2000

2500

3000

3500

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9

Seconds

Packet Rate (pps) Packet Loss (packets)

Path D - spf delay = 200 ms

0

500

1000

1500

2000

2500

3000

3500

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9

Seconds

Packet Rate (pps) Packet Loss (packets)

Path D - spf delay = 50 ms

0

500

1000

1500

2000

2500

3000

3500

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9

Seconds

Packet Rate (pps) Packet Loss (packets)

Path E - spf delay = 200 ms.

0

500

1000

1500

2000

2500

3000

3500

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9

Packet Rate (pps) Packet Loss (packets)

Path E - spf delay = 50 ms

0

500

1000

1500

2000

2500

3000

3500

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9

Packet Rate (pps) Packet Loss (packets)

Path F - spf delay = 200 ms.

0

500

1000

1500

2000

2500

3000

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9

Packet Rate (pps) Packet Loss (packets)

Path F - spf delay = 50 ms.

0

500

1000

1500

2000

2500

3000

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9

Packet Rate (pps) Packet Loss (packets)

Observations (so far…)

• Decreasing spf delay seems to make convergence time worse

• MPLS FRR offers a possible alternative way to minimize convergence time

• More analysis needed

• More testing needed