1

Lecture 17

Fairness and Neutrality

University of Nevada – RenoComputer Science & Engineering Department

Fall 2015

CS 791 Special Topics:Network Architectures and Economics

2

Outline Neutrality

Definition of net neutrality Key issues

3

What is Net Neutrality?“Net neutrality is a principle that advocates no

restrictions by Internet service providers or governments on consumers' access to networks

that participate in the Internet. Specifically, network neutrality would prevent restrictions on content, sites, platforms, types of equipment that may be

attached, and modes of communication.”-- Wikipedia

Confusion on equality Equal service to all packets? Equal service to everyone? Equal service to everyone within the same class? …

4

Differentiation vs. Discrimination Differentiation:

Giving priority to some traffic Is this bad? Under what conditions?

Discrimination: Why is it different than differentiation? Explicit blocking of traffic based on what?

5

The Current Internet Market

Current stake holders: Users Content providers

• Google, Yahoo!, Time Warner Access providers

• Time Warner, Charter Carrier providers

• AT&T, Sprint, Qwest

Revenue sharing among providers: content >> access > carrier e.g., 90% >> 7% > 3% (speculative

numbers)

6

The Debate

Should we impose more regulation or not?

What are the current regulations? Very traditional regulations that exist in any

other sector. no monopolies

• AT&T was forced to split several times before no vertical integration

• a content provider should not be an access provider

have to obey the contracts• cannot under-deliver in regards to what you

promised

7

The Debate Issues

Openness• Make it available to everyone!

Cost• It is costlier to engineer, just overprovision and let

everyone get the same performance as those elite would get.

Application requirements• More stringent QoS requirements are being asked,

e.g. convergence of IPTV. Business models

• What is the revenue sharing and end-to-end pricing to keep private-sector motivation high?

8

Openness How open is it?

Real-time vs. Best-effort• VoIP – a 911 call?• Peer-to-peer traffic – an unlicensed video

download?

Security and Privacy• Active filtering of “disliked” content – e.g.

parental controls• Blocking of ports – e.g. FTP port

What is the right level of openness necessary so that application-specific requirements are met while

keeping the doors for innovations?

9

Cost Capacity costs

How much more capacity is needed to keep up with the application requirements?

What is the extra capacity needed to match the performance of a premium quality service?

Deployment costs Which one is easier to deploy? By how

much?

10

Thick (Over-provisioned) or Thin (Engineered) Pipes?

Thin: How to deal with bursts/overload?And meet premium SLAs… !

Thick: Cost of overprovisioning?Can this commodity model break even?

0 40000 80000

10000

0

rate

time

[Jim Roberts et al.]

Media-rich applications require performance guarantees:

e.g.: VoIP requires <300ms round-trip delay, <1% loss

How to respond to these application needs?

CoS approach: provide priority (i.e. higher class) to premium traffic

Classless (best-effort) service approach: over-provision the capacity

Question: How much extra capacity does the classless service require to match the performance of the higher class (premium) service in the CoS approach?

0 40000 80000

10000

0

rate

time

11

Two Approaches: CoS vs. Classless

Premium

BE

D

CoS Link (differentiated)

D

Prem= gD

BE=(1-g)D

D

• GIVEN: D, D and a performance target (i.e. ttarget or ptarget)

• FIND: The minimum N that gives the same performance as in the premium class of the CoS case?

N=?

Classless Link (neutral)

BE

Sch

edulin

g(e

.g. pri

ori

ty)

12

REC: Required Extra Capacity

REC = <required neutral link capacity> - <CoS link

capacity>= N - D (rate)= 100(N/D – 1) (%)

How to quantify REC?

13

Analytical Link Model: Poisson traffic• Assume:

– Poisson traffic, Exponential packet lengths for traffic in each class i.e.• Premium class traffic is Poisson with g D

• Best-effort class traffic is Poisson with (1-g) D– The aggregate traffic for the neutral link is also Poisson with

rate D

Both the performance target and the REC can be expressed in terms of two key parameters: (i) ρ – utilization, (ii) g – proportion of premium traffic.

14

More Bursty Traffic: MMPP• MMPP = Markov-Modulated Poisson

Process– Easy to do the math…– Simplest MMPP is of two states.

• MMPP traffic with mean D

– Traffic w/ equivalent rate to the neutral case, but w/ more burstiness.

1 2

aar

aaar

ar

1

1

2

1

Higher r means more bursty traffic.

15Simulated Link Model: DelayMMPP/M/1 model

a=0.5, r=4If packet size is 1KB and the CoS link is D = 1Gb/s:5,000packets of delay = 40.1ms

Surface color shows the performance target.

REC can be quite high even for very small g and medium utilizations.

16Simulated Link Model: DelayMMPP/M/1 model

a=0.5, r=4

REC increases as link utilization increases

REC is large even for small proportion of premium traffic

Can be drawn in multiple 2-d graphs

17

Simulated Link Model: LossMMPP/M/1/K model

The graphs are generic for various buffer sizes. An example: For a 10Mb/s link carrying 1KB packets:

K = ~15pkts 25ms buffer time

K = ~60pkts 100ms buffer time

REC for the same performance target decreases as buffer size increases

18

Simulated Link Model: LRD Traffic

Internet traffic : known to be LRD with Hurst parameter value between 0.7 and 0.9. REC for Hurst=0.75 is significantly higher than our 2-state MMPP model results. We also observed that REC increases as Hurst value increases towards 0.9.

DELAY – LRD/D/1 LOSS – LRD/D/1/K

Also looked at closed-loop traffic - many TCP flows - and observed similar trends. We further looked at the case when Premium traffic is CBR and BE is TCP, and this increased

REC further.

19

NREC: Two ways to calculate• Calculate network REC (NREC):

We calculated NRECs for the Rocketfuel topologies: – Used the MMPP link model (a=0.5 and r=4) or the LRD link model

(H=0.75) – Much more conservative than real or TCP traffic– Assumed K=100ms buffer time– Only report Sprintlink, as the other topologies gave higher REC

values

total extra capacity needed on the whole network

average extra capacity needed on each link

20

NREC for Sprintlink: G2G Delay

Solid lines are NRECI and dashed lines are NRECA

NREC can be much higher than 100% for a network operating with 60% utilization.

10ms queueing delay target for VoIP may require large REC values.

21

NREC for Sprintlink: G2G Loss

Solid lines are NRECI and dashed lines are NRECA

NREC can be much higher than 1000% even for a network operating with 40% utilization.

0.1% loss target may require large REC values.

22

Capacity Costs: Summary• A framework to study REC for delay or loss being the performance

target.• Link model

– REC grows when:• traffic becomes more bursty• the utilization of the CoS link becomes higher• the performance target becomes tighter• the fraction g of the Premium class traffic becomes smaller

– Closed-loop (e.g., TCP) or LRD traffic further increases REC• Network model:

– For legacy g2g performance targets, REC ranges from 50% to over 100% as g reduces below 0.5 and the utilization goes up to 60%.

• Future trends/work:– The performance targets will keep becoming tighter. REC is

high perpetually – not just today, but in future also.. – The value of g is a crucial factor. Small g does not

necessary favor a classless network.– Further research should estimate the actual costs of CoS

and classless designs, as scheduling & management complexity need to be considered.

23

Application Requirements and IP Convergence

IP is cheaper than any other dedicated networking medium! An attractive banana for convergence VoIP, Video Conferencing, Gaming IPTV, Video-on-Demand

What is next? Cloud Computing? Network Storage? ..

24

Internet Industry’s Challenge… Provide cost-effective Internet access while:

Applications continue to diversify relative to demands on the network

The number of users continues to grow at an annual rate of 12.5%

The average traffic per user is expected to grow even faster at 25% per year

APPLICATION MAXIMUM ONE-WAY DELAY

PACKET LOSS IN THE NETWORK

IPTV <100 msec <0.01% Video-on-Demand <50 msec <0.001% VoIP <150 msec <0.1% Video Conferencing <150 msec <0.05% Gaming <50 msec <0.1%

25

Application Requirements and IP Convergence

Convergence is bringing tighter application requirements

How to respond to these diversifying and growing application demands while keeping

the service provisioning as a viable and sustainable business?

26

Some proposed legislation would force the service provider to participate only in the rightmost side of the continuum.

Advertising Driven Consumer Paid

The Internet needs the ability to support different business models depending on the application or user to optimize the consumer value and continue investment in broadband infrastructure.

• AOL is moving away from a subscription based model to an advertising model.

• NYT is diversifying its revenue by growing its online edition which will be primarily advertisement funded.

A blended mix of cost assignment between end user and content owner would serve the end user’s needs.

Business Model Flexibility is Critical

27

Lecture 17: Summary Openness: differentiation vs.

discrimination

Capacity vs. dollar costs

Required Extra Capacity

IP convergence is inevitable

The right business model

28

Lecture 17: Reading Network Neutrality, Wikipedia,

http://en.wikipedia.org/wiki/Network_neutrality

Clark, Network Neutrality: Words of Power and 800-Pound Gorillas, International Journal of Communication, 2007.