Pay or Perish: Decoding the Economic Gravity of Premium Peering

TL;DR

In the evolving Internet, "Best Effort" is no longer enough for giants like Netflix. This paper provides a mathematical blueprint for why CPs choose to pay APs for direct access. By introducing the Intrinsic Value of Premium Peering (VoPP), it proves that peering is not just a technical choice, but a strategic defense against "peer pressure" and user churn.

The "Broken" Settlement Model

The Internet was built on the gentleman's agreement of settlement-free peering. But when Netflix began consuming 34% of downstream traffic, the status quo shattered. APs (like Comcast) began demanding payment from CPs. The FCC exempted these deals from Net Neutrality rules due to a lack of "background." This paper fills that void, moving beyond simple traffic ratios to analyze the Economic Choice Model of users.

Methodology: The Complementary Choice Model

The author rejects the idea that CPs and APs are independent. Instead, they are Complementary Services. A user needs both to watch a video.

The Core Equation: User Stickiness

The model assumes users have "stickiness" ($\alpha$ for CPs, $\beta$ for APs). If a peering link is slow (θ=0), whether a user leaves depends on how much they love the content versus how many other ISPs they can switch to.

Figure 1: The Weighting Factor (ρ) that determines how market shares shift when a peering agreement is established or terminated.

The "VoPP" (Value of Premium Peering)

The paper defines an Intrinsic VoPP ($v_i$): $v_i=(1-{\alpha }_i{\delta }_i)q_i$

• $q_i$: The per-user value of the content.
• ${\delta }_i$: Preservation factor (how much value remains if the connection is slow).
• ${\alpha }_i$: User stickiness to the CP.

If the AP’s price ($p$) is higher than $v_i$, the CP will refuse to peer. If $p\le v_i$, peering is a go.

Strategic Insights: The High-Value CP Dilemma

One of the paper's most fascinating findings is the concept of Peer Pressure.

1. High-Value CPs (Netflix): They peer because if they don't, and a low-value CP does, the high-value CP loses its "elastic" (non-sticky) users to the better-performing competitor.
2. Low-Value CPs: They behave like "contrarians." They peer only if the big players haven't already saturated the quality expectation at that price point.

Figure 2: Analysis of Nash Equilibria. Notice how the peering strategy of one CP (ϑ₁) is fundamentally altered by the price and decision of the other.

Critical Insight: Market Baseline vs. Stickiness

• AP Baseline Share: Surprisingly, an AP's current size ($\psi$) matters less than the stickiness ($\beta$) of its users. An AP with a 5% market share but 99% sticky users can still act like a "local monopoly" and demand high peering fees.
• CP Baseline Share: For CPs ($\phi$), size matters more. A larger CP has more "to lose" in terms of non-sticky users, driving them into peering agreements faster than smaller niche providers.

Conclusion and Future Outlook

The paper concludes that transparency—making it easier for users to switch ISPs—is a more effective regulatory tool than capping peering prices. If $\beta$ (AP stickiness) decreases, APs are forced into a "price war," naturally lowering the premium peering costs for CPs.

Takeaway: In the "Pay or Perish" world of the modern web, your economic value is determined not by how much traffic you send, but by how easily your users can leave you.