Understanding Partial Reachability in the Internet Core

TL;DR

This work redefines the Internet core as the strongly-connected component comprising more than $50\%$ of active public IP addresses that can bidirectionally reach one another, enabling a measurement-based, authority-free notion of a global core. It introduces two detectors, Taitao for peninsulas and Chiloe for islands, and applies them to large-scale data from Trinocular and RIPE Atlas, with CAIDA Ark validation, demonstrating that peninsulas and islands are pervasive and often more frequent than outages. The approach improves DNSmon sensitivity, clarifies outage reports, and provides a framework for policy discussions on fragmentation, sovereignty, and peering disputes. Overall, the results show persistent partial reachability is a fundamental Internet property with substantial operational and governance implications, motivating better measurement practices and policy-aware infrastructure design.

Abstract

Routing strives to connect all the Internet, but compete: political pressure threatens routing fragmentation; architectural changes such as private clouds, carrier-grade NAT, and firewalls make connectivity conditional; and commercial disputes create partial reachability for days or years. This paper suggests *persistent, partial reachability is fundamental to the Internet* and an underexplored problem. We first *derive a conceptual definition of the Internet core* based on connectivity, not authority. We identify *peninsulas*: persistent, partial connectivity; and *islands*: when computers are partitioned from the Internet core. Second, we develop algorithms to observe each across the Internet, and apply them to two existing measurement systems: Trinocular, where 6 locations observe 5M networks frequently, and RIPE Atlas, where 13k locations scan the DNS roots frequently. Cross-validation shows our findings are stable over *three years of data*, and consistent with as few as 3 geographically-distributed observers. We validate peninsulas and islands against CAIDA Ark, showing good recall (0.94) and bounding precision between 0.42 and 0.82. Finally, our work has broad practical impact: we show that *peninsulas are more common than Internet outages*. Factoring out peninsulas and islands as noise can *improve existing measurement systems*; their ``noise'' is $5\times$ to $9.7\times$ larger than the operational events in RIPE's DNSmon. We show that most peninsula events are routing transients (45\%), but most peninsula-time (90\%) is due to a few (7\%) long-lived events. Our work helps inform Internet policy and governance, with our neutral definition showing no single country or organization can unilaterally control the Internet core.

Figures (18)

• Figure 1: $A$, $B$ and $C$ are the connected core, $B$ and $C$ peninsulas, $D$ and $E$ islands, $X$ is out.
• Figure 2: Distribution of block-time fraction: all-down (left), disagreement (center), and all-up (right), events $\ge 1$ hour. Data: 3.7M blocks, 2017-10-06 to -11-16, A30.
• Figure 3: Peninsulas measured with per-site down events longer than 5 hours. Dataset A30, 2017q4.
• Figure 4: Cumulative peninsulas and peninsula duration. Dataset A30, 2017q4.
• Figure 5: CDF of islands detected by Chiloe for data from Trinocular (3 years, Datasets A28-A39) and Atlas (2021q3).