Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Reasoning About Internet Connectivity

Guillermo Baltra, Tarang Saluja, Yuri Pradkin, John Heidemann

TL;DR

The paper defines the Internet core as the strongly connected component of more than $0.5$ of active public IP addresses that can initiate communication with each other, yielding separate IPv4 and IPv6 cores. It argues that partial reachability manifests as peninsulas and islands, shaping policy, architecture, and operation. Using operational data sources, it demonstrates how this core concept improves measurement sensitivity (e.g., DNSmon) by distinguishing persistent routing issues from misconfigurations. The work concludes that no single country or organization can unilaterally control the Internet core, underscoring global interdependence and providing a principled framework for assessing fragmentation risks and resilience.

Abstract

Innovation in the Internet requires a global Internet core to enable communication between users in ISPs and services in the cloud. Today, this Internet core is challenged by partial reachability: political pressure threatens fragmentation by nationality, architectural changes such as carrier-grade NAT make connectivity conditional, and operational problems and commercial disputes make reachability incomplete for months. We assert that partial reachability is a fundamental part of the Internet core. While other studies address partial reachability, this paper is the first to provide a conceptual definition of the Internet core so we can reason about reachability from principles first. Following the Internet design, our definition is guided by reachability, not authority. Its corollaries are peninsulas: persistent regions of partial connectivity; and islands: when networks are partitioned from the Internet core. We show that the concept of peninsulas and islands can improve existing measurement systems. In one example, they show that RIPE's DNSmon suffers misconfiguration and persistent network problems that are important, but risk obscuring operationally important connectivity changes because they are $5\times$ to $9.7\times$ larger. Our evaluation also informs policy questions, showing no single country or organization can unilaterally control the Internet core.

Reasoning About Internet Connectivity

TL;DR

The paper defines the Internet core as the strongly connected component of more than of active public IP addresses that can initiate communication with each other, yielding separate IPv4 and IPv6 cores. It argues that partial reachability manifests as peninsulas and islands, shaping policy, architecture, and operation. Using operational data sources, it demonstrates how this core concept improves measurement sensitivity (e.g., DNSmon) by distinguishing persistent routing issues from misconfigurations. The work concludes that no single country or organization can unilaterally control the Internet core, underscoring global interdependence and providing a principled framework for assessing fragmentation risks and resilience.

Abstract

Innovation in the Internet requires a global Internet core to enable communication between users in ISPs and services in the cloud. Today, this Internet core is challenged by partial reachability: political pressure threatens fragmentation by nationality, architectural changes such as carrier-grade NAT make connectivity conditional, and operational problems and commercial disputes make reachability incomplete for months. We assert that partial reachability is a fundamental part of the Internet core. While other studies address partial reachability, this paper is the first to provide a conceptual definition of the Internet core so we can reason about reachability from principles first. Following the Internet design, our definition is guided by reachability, not authority. Its corollaries are peninsulas: persistent regions of partial connectivity; and islands: when networks are partitioned from the Internet core. We show that the concept of peninsulas and islands can improve existing measurement systems. In one example, they show that RIPE's DNSmon suffers misconfiguration and persistent network problems that are important, but risk obscuring operationally important connectivity changes because they are to larger. Our evaluation also informs policy questions, showing no single country or organization can unilaterally control the Internet core.
Paper Structure (18 sections, 6 figures)

This paper contains 18 sections, 6 figures.

Table of Contents

  1. Introduction
  2. Problem: Partial Reachability
  3. Defining the Internet Core
  4. Away from the Core: Unreachability
  5. Outages
  6. Islands: Isolated Networks
  7. Peninsulas: Partial Connectivity
  8. Applying the Definition
  9. Resolving Conflicting Claims
  10. Policy Applications of the Definition
  11. Architecture and Operation Application
  12. Can the Internet Core Partition?
  13. Definitions Clarify DNSmon Sensitivity
  14. Related Work
  15. Conclusions
  16. ...and 3 more sections

Figures (6)

  • Figure 1: Blocks $A$, $B$ and $C$ are the connected core, with $B$ and $C$ peninsulas; $D$ and $E$ islands; $X$ an outage.
  • Figure 2: RIR IPv4 hosts and IPv6 /32 alloc. iana_v4iana_v6.
  • Figure 3: Fraction of VPs observing islands and peninsulas for IPv4 and IPv6 during 2022.
  • Figure 4: Atlas queries from all available VPs to 13 Root Servers for IPv4 and IPv6 on 2022-07-23.
  • Figure 5: Estimates of an island in Trinocular data starting 2017-06-03t23:06Z and lasting just longer than one hour.
  • ...and 1 more figures