Peering vs. Transit Adnan Ahmed University of Iowa University of - - PowerPoint PPT Presentation

▶
peering vs transit
SMART_READER_LITE
LIVE PREVIEW

Peering vs. Transit Adnan Ahmed University of Iowa University of - - PowerPoint PPT Presentation

Dec 12, 2023 230 likes •616 views

Peering vs. Transit Adnan Ahmed University of Iowa University of Iowa Introduction Transit Provides connectivity to the Internet Traffic volume based fees Peering Bilateral exchange Settlement-free (no fee) University

slide-1
SLIDE 1

Peering vs. Transit

Adnan Ahmed University of Iowa

University of Iowa

slide-2
SLIDE 2

Introduction

  • Transit
  • Provides connectivity to

the Internet

  • Traffic volume based fees
  • Peering
  • Bilateral exchange
  • Settlement-free (no fee)

University of Iowa

  • 1. Introduction
slide-3
SLIDE 3

Related work

Interconnection strategies in peering ecosystem

  • Agent-based analysis [Lodhi-Dhamdhere, SIGMETRICS ’12]
  • Open-peering [Lodhi et al., Infocom ’14]
  • Game-theoretic models [Accongiagioco et al., IFIP

’14][Badasyan-Chakrabarti, Telecommunications Policy ’08]

  • Complexities in decision making [Lodhi et al., Infocom ‘15]

Evolution of peering and topological Impact

  • Network model [Dhamdhere-Dovrolis, CoNEXT ’10]
  • Remote peering [Castro et al., CoNEXT ‘14]
  • IXP study [Ager et al., SIGCOMM ‘12]
  • PeeringDB analysis [Lodhi et al., SIGCOMM CCR ’14]

Our focus

  • Large-scale measurement based performance comparison

University of Iowa

  • 2. Related work
slide-4
SLIDE 4

Methodology

  • Throughput measurements
  • Strain the network
  • Delay measurements using ICMP packet probing
  • Rate limiting at ISPs
  • Our approach
  • HTTP based end-to-end delay measurements

University of Iowa

  • 3. Methodology
slide-5
SLIDE 5

The big picture

University of Iowa

  • 4. The big picture
slide-6
SLIDE 6

The big picture

University of Iowa

  • 4. The big picture
slide-7
SLIDE 7

The big picture

University of Iowa

  • 4. The big picture
slide-8
SLIDE 8

Our approach

University of Iowa

  • 5. Our approach
slide-9
SLIDE 9

Our approach

University of Iowa

  • 5. Our approach
slide-10
SLIDE 10

Our approach

University of Iowa

  • 5. Our approach
slide-11
SLIDE 11

Our approach

University of Iowa

  • 5. Our approach
slide-12
SLIDE 12

Our approach

University of Iowa

  • 5. Our approach
slide-13
SLIDE 13

Data collection

University of Iowa

  • 6. Data collection
  • A commercial CDN
  • Collected across PoPs at 19 IXPs
  • 1M measurements
  • ~350K clients
  • 360 Autonomous systems
slide-14
SLIDE 14

Peering vs Transit

University of Iowa

  • 7. Peering vs Transit
slide-15
SLIDE 15

Peering vs Transit

University of Iowa

  • 7. Peering vs Transit
slide-16
SLIDE 16

Peering vs Transit

University of Iowa

  • 7. Peering vs Transit
slide-17
SLIDE 17

Peering vs Transit

University of Iowa

  • 7. Peering vs Transit
slide-18
SLIDE 18

RTT components

University of Iowa

  • 8. RTT components
slide-19
SLIDE 19

RTT components

University of Iowa

  • 8. RTT components
  • Transmission ~ 0
  • small size of pixel tag
slide-20
SLIDE 20

RTT components

University of Iowa

  • 8. RTT components
  • Transmission ~ 0
  • small size of pixel tag
  • RTT ~ Propagation delay + Queueing delay
slide-21
SLIDE 21

RTT components

University of Iowa

  • 8. RTT components

RTT measurements over time

slide-22
SLIDE 22

RTT components

University of Iowa

  • 8. RTT components

RTT measurements over time

  • Diurnal pattern
slide-23
SLIDE 23

RTT components

University of Iowa

  • 8. RTT components
  • Diurnal pattern
  • Time series latency pings
slide-24
SLIDE 24

RTT components

University of Iowa

  • 8. RTT components
  • Diurnal pattern
  • Time series latency pings
slide-25
SLIDE 25

RTT components

University of Iowa

  • 8. RTT components
  • Diurnal pattern
  • Time series latency pings
  • Propagation delay < RTTmin
slide-26
SLIDE 26

RTT components

University of Iowa

  • 8. RTT components
  • Diurnal pattern
  • Time series latency pings
  • Propagation delay < RTTmin
slide-27
SLIDE 27

RTT components

University of Iowa

  • 8. RTT components
  • Diurnal pattern
  • Time series latency pings
  • Propagation delay < RTTmin
slide-28
SLIDE 28

RTT components

University of Iowa

  • 8. RTT components
  • Diurnal pattern
  • Time series latency pings
  • Propagation delay < RTTmin
  • Maximum queueing delay

~ RTTmax - RTTmin

slide-29
SLIDE 29

Propagation and Queueing delays

University of Iowa

  • 9. Propagation and Queueing delays
slide-30
SLIDE 30

Propagation and Queueing delays

University of Iowa

  • 9. Propagation and Queueing delays

Peering is better

slide-31
SLIDE 31

Propagation and Queueing delays

University of Iowa

  • 9. Propagation and Queueing delays

Transit is better

slide-32
SLIDE 32

Propagation and Queueing delays

University of Iowa

  • 9. Propagation and Queueing delays

I IV II III

slide-33
SLIDE 33

Propagation and Queueing delays

University of Iowa

  • 9. Propagation and Queueing delays

I IV II III

slide-34
SLIDE 34

Propagation and Queueing delays

University of Iowa

  • 9. Propagation and Queueing delays

I IV II III

slide-35
SLIDE 35

Path length comparison

University of Iowa

  • 10. Path length comparison
slide-36
SLIDE 36

Path length comparison

University of Iowa

  • 10. Path length comparison
slide-37
SLIDE 37

Path length comparison

University of Iowa

  • 10. Path length comparison

Shorter path length via peering

slide-38
SLIDE 38

Conclusion

  • Peering generally outperforms transit for a majority
  • f clients
  • Peering almost always has better propagation

delays

  • Shorter path lengths for peering
  • Transit sometimes has better queueing delays
  • Under-provisioned peering paths

University of Iowa

  • 11. Conclusion