Interconnection measurement for policy and for research Scott Jordan - - PowerPoint PPT Presentation

interconnection measurement for policy and for research
SMART_READER_LITE
LIVE PREVIEW

Interconnection measurement for policy and for research Scott Jordan - - PowerPoint PPT Presentation

Interconnection measurement for policy and for research Scott Jordan University of California, Irvine Is interconnection a policy issue? Aspiration? reach unblocked choice Legal imperative? Common carrier: duty


slide-1
SLIDE 1

Interconnection measurement for policy and for research

Scott Jordan University of California, Irvine

slide-2
SLIDE 2

Is interconnection a policy issue?

  • Aspiration?

reach

unblocked

choice

  • Legal imperative?

Common carrier:

 duty to interconnect  prohibition on unreasonable or unreasonably discriminatory charges and practices

RIF:

 no legal authority in Comm Act

 no current harms  market will mitigate risk of blocking, throttling, prioritization using interconnection

 legal authority through anti‐trust, consumer protection

 anti‐competitive conduct

Interconnection / Scott Jordan / University of California Irvine

2

slide-3
SLIDE 3

Policy goals

  • prohibit:

 harmful discriminatory interconnection practices  harmful refusal to upgrade interconnection

capacity

 unreasonable access fees   rules, oversight

  • discourage (rather than prohibit):

 (the stuff above)   transparency

Interconnection / Scott Jordan / University of California Irvine

3

slide-4
SLIDE 4

What to measure

  • performance metrics:

 border router to border router, across an IXP:

 utilization  latency  packet loss

Interconnection / Scott Jordan / University of California Irvine

4

slide-5
SLIDE 5

Research goals

  • develop model for researchers:

 what determines the cost of interconnection?  what determines the value of interconnection?  how should the price of interconnection be

determined?

 who should pay for interconnection?

  • develop model for oversight:

 is an offered interconnection arrangement

unreasonable, or unreasonably discriminatory?

Interconnection / Scott Jordan / University of California Irvine

5

slide-6
SLIDE 6

What to measure

  • traffic matrices:

 source IXP to destination IXP

 content provider  IXP at which traffic enters transit provider’s network (if any)  IXP at which traffic enters ISP’s network  closest IXP to the customer

Interconnection / Scott Jordan / University of California Irvine

6

slide-7
SLIDE 7

What to measure

  • one‐way delay matrices:

 source IXP to destination IXP

 border router in content provider’s network  through transit provider (if any)  border router in ISP’s network

Interconnection / Scott Jordan / University of California Irvine

7

slide-8
SLIDE 8

Who should measure

  • ISPs

 e.g. transparency requirement

  • FCC

 e.g. Measuring Broadband America

  • Researchers

 usually from the edge

Interconnection / Scott Jordan / University of California Irvine

8

slide-9
SLIDE 9

Who should measure what?

  • border router to border router

 ISP is in the best position to measure  MBA could work with ISPs  measurement from the edge is possible

Interconnection / Scott Jordan / University of California Irvine

9

slide-10
SLIDE 10

Who should measure what?

  • direct connection between content provider

and ISP

 ISP is in the best position to measure  more difficult for MBA (route across ISP network)  hoping measurement from the edge is possible

Interconnection / Scott Jordan / University of California Irvine

10

slide-11
SLIDE 11

Who should measure what?

  • indirect connection through a transit provider

 ISP could work with transit provider  more difficult for MBA [need to work with ISP &

transit provider (?)]

 hoping measurement from the edge is possible

Interconnection / Scott Jordan / University of California Irvine

11