From Peering Economics to Interconnection Sovereignty
A research note examining new global evidence on the economics of Internet Exchange Points and connecting it to carrier neutrality, local economic development, traffic localization, regional resilience, and an emerging concept of interconnection sovereignty.
Internet Exchange Points as Local Economic and Civic Infrastructure
Executive Summary
A new Internet Society study, Dissecting the Cost of Peering: Evidence from Global Data on Internet Exchange Points (IXPs), provides quantitative evidence for an argument that has been emerging throughout my work on carrier-neutral Internet exchanges: where networks interconnect matters economically, operationally, and institutionally.
The study compares peering and IP-transit costs across multiple countries and cities and finds peering substantially less expensive on average. Greater savings from peering are associated with lower broadband prices, greater mobile coverage, and higher aggregate telecommunications revenues. Just as importantly, the researchers find that outcomes depend on peering intensity, not simply whether an Internet Exchange Point exists.
That distinction connects directly with my work on a carrier-neutral exchange in Michigan’s Eastern Upper Peninsula. In my earlier planning, I have treated neutrality as more than an engineering characteristic. A carrier-neutral exchange creates a place where participating networks can interconnect without allowing a single carrier to determine the conditions of exchange.
Taken together, these ideas suggest a broader proposition:
An Internet Exchange Point is not simply a place where networks exchange packets. It is an institution through which a region can retain more of the economic, operational, and decision-making value produced by its connectivity.
This provides a bridge between three strands of my work that are often discussed separately: carrier neutrality, local economic development, and local or tribal sovereignty.
From Access to Interconnection
Most broadband policy begins at the edge of the network: can a household, business, school, or institution obtain an adequate connection? That question remains essential, but it is incomplete.
Once access networks exist, another question follows:
Where does the traffic go?
Without local interconnection, traffic between nearby organizations may travel through distant upstream networks before returning to the region. An Internet Exchange Point changes that relationship by giving independently operated networks a place to exchange appropriate traffic directly. Tagat and Phokeer describe precisely this mechanism: IXPs can allow traffic to remain local rather than relying solely on distant third-party transit, with potential benefits for cost, performance, resilience, and the strength of the local Internet ecosystem.
That is remarkably close to the infrastructure strategy I have been developing for the Eastern Upper Peninsula. My earlier planning has focused on carrier-neutral fiber meet points and an international Internet exchange, while recognizing Sault Ste. Marie’s position on the Canadian border as an opportunity for additional international connectivity and routing diversity.
The new research therefore helps move the discussion one layer deeper:
Broadband access connects a community to the Internet. Interconnection determines how that community participates in the Internet.
Carrier Neutrality as Economic Infrastructure
Carrier neutrality can sometimes sound like an architectural preference. The Internet Society research strengthens the case for treating it as an economic-development principle.
Tagat and Phokeer recommend neutral, open, voluntary, transparent, and commercially sustainable IXP ecosystems, with participation from diverse networks, including Internet service providers, content providers, and government networks. Their policy recommendation is not that governments manage individual peering relationships, but that policymakers help establish the conditions within which healthy interconnection ecosystems can develop.
This aligns strongly with my carrier-neutral IXP work for the Eastern Upper Peninsula, in which neutrality is intended to prevent a single network from controlling the exchange and to preserve opportunities for multiple providers to participate on equitable terms.
It also aligns with research I have collected on distributed and neutral exchanges as catalysts for digital ecosystem development. In that model, neutral interconnection can create redundancy and choice while allowing data centers, carriers, content providers, and other infrastructure participants to develop around the exchange.
The emerging economic logic is:
neutral interconnection → broader participation → greater peering intensity → lower interconnection friction → stronger digital ecosystem
The Internet Society paper does not establish every link in that chain as causal. In fact, the authors are careful to acknowledge the limitations of their models. But the study provides empirical support for several important relationships within the chain.
An IXP as an Economic Retention Mechanism
The most interesting connection to my broader regional work may be economic retention.
Tagat and Phokeer’s dataset produces an average observed peering cost of approximately $0.054 per Mbps per month, compared with approximately $18.27 per Mbps per month for IP transit. Those numbers require substantial caution. The authors note that their transit observations frequently represent international or backbone transit, while domestic wholesale prices can be much lower. They also acknowledge that IXP port fees do not represent the entire cost of peering. Their survey of IXP operators found that port charges accounted for approximately 44 percent of reported peering-related costs on average.
For that reason, it would be inappropriate to apply the headline price differential directly to the Eastern Upper Peninsula and claim a corresponding regional savings figure.
The stronger conclusion is structural. Where enough traffic can be exchanged locally, networks have an opportunity to reduce their dependence on purchased upstream transit. The study associates greater peering savings with lower broadband prices, expanded mobile network coverage, and higher total telecommunications revenues, although the authors appropriately caution against interpreting every observed relationship as a universal causal effect.
This creates a useful way of thinking about an IXP:
An Internet Exchange Point can function as a regional economic retention mechanism.
Instead of purchasing every unit of connectivity from somewhere else, participating networks can exchange some traffic through a shared local marketplace.
This complements an argument that runs throughout my broader regional work: infrastructure becomes more valuable when it increases the capability of the community around it rather than merely delivering a service to it.
Interconnection extends that principle from the access network into the underlying economic structure of the Internet.
From Data Sovereignty to Interconnection Sovereignty
This is where I think the Internet Society research opens an especially important extension of my existing work.
My work around digital opportunity has treated tribal data sovereignty as an essential consideration in how information about Indigenous communities is collected, owned, governed, interpreted, and applied. Connectivity alone does not establish sovereignty over information.
Other work on data-center and interconnection ecosystems introduces another dimension: sovereign data flows. Organizations increasingly care not only about where data is stored, but about which jurisdictions, infrastructure, intermediaries, and network paths their information traverses.
The Internet Society study contributes a third component: traffic localization. Tagat and Phokeer explicitly examine the degree to which content can be reached locally rather than through international transit and consider localization as one of the forces associated with peering.
Together, these ideas suggest another concept that deserves further exploration: interconnection sovereignty.
I use interconnection sovereignty here as my own synthesis rather than as a concept established by the Internet Society study. It describes the practical ability of a community, institution, tribal nation, or region to exercise meaningful choices over its network relationships, traffic paths, interconnection locations, and dependence upon external intermediaries.
This creates at least three related layers of digital sovereignty:
| Layer | Central Question |
|---|---|
| Connectivity sovereignty | Can we connect on reasonable, resilient, and competitive terms? |
| Data sovereignty | Who governs the collection, ownership, storage, use, and interpretation of our data? |
| Interconnection sovereignty | Who determines how our networks connect, where our traffic travels, and which intermediaries we depend upon? |
These concepts should not be collapsed into one another. Local infrastructure does not automatically create sovereignty. Community ownership does not automatically create carrier neutrality. An Internet Exchange Point does not guarantee control over data or traffic.
But neutral local interconnection creates options, and meaningful options are one of the prerequisites for greater agency.
Why the Eastern Upper Peninsula Is an Interesting Case
The Eastern Upper Peninsula presents an especially interesting environment in which to explore these ideas because the region combines rurality, an international border, tribal nations, island communities, terrestrial broadband, satellite infrastructure, energy infrastructure, educational networks, and an emerging technology and workforce-development ecosystem.
My regional infrastructure work has already connected the proposed neutral IXP with middle-mile infrastructure, earth-ground-station interconnection, routing diversity, network resilience, emergency communications, island connectivity, and potential international connections toward Canada.
The Internet Society research helps place an economic mechanism underneath that strategy.
The regional proposition begins to look like this:
Last-mile fiber creates access.
Middle-mile infrastructure creates reach.
An Internet Exchange Point creates a local interconnection market.
Carrier neutrality preserves choice within that market.
Local peering creates opportunities to retain traffic and some of its associated economic activity locally.
Cross-border interconnection creates additional routes, relationships, and potential markets.
Regional institutions and tribal nations gain additional choices regarding how their networks participate in the larger Internet ecosystem.
That is a substantially richer infrastructure argument than simply saying that an IXP can reduce latency.
Participation Matters More Than the Building
One of the most important implications of Dissecting the Cost of Peering is that constructing an Internet Exchange Point is not enough. Tagat and Phokeer emphasize peering intensity, essentially the degree to which networks actually participate in an exchange ecosystem.
That distinction has immediate implications for the Eastern Upper Peninsula effort.
The objective should not simply be:
Build an IXP.
It should be:
Build an interconnection ecosystem.
That means developing participation among regional fiber operators, Internet service providers, mobile providers, tribal networks, educational and research networks, governments, universities and colleges, cloud and content providers, content delivery networks, satellite operators, and other autonomous networks.
This is consistent with how the EUP initiative has already begun developing. The regional conversation has brought together institutions including Connected Nation, the Sault Ste. Marie Tribe of Chippewa Indians, the Chippewa County Economic Development Corporation, Eastern Upper Peninsula Intermediate School District, Cloverland Electric Cooperative, Bay Mills Community College, and Lake Superior State University.
Seen through the findings of the new research, this coalition-building is not peripheral stakeholder engagement.
Participation is part of the mechanism through which an IXP produces value.
Interconnection and the Infrastructure Dividend
There is also an important connection between this work and my emerging thinking about an infrastructure dividend.
As data centers, artificial intelligence infrastructure, fiber networks, energy systems, and other large-scale infrastructure become increasingly important to economic development, communities face a fundamental question: how do they participate in the value created by infrastructure rather than simply hosting it or purchasing services from it?
An Internet Exchange Point offers one practical example of how infrastructure can change the economics surrounding it.
The exchange itself does not have to capture all the resulting value. Instead, it can affect the ecosystem around it. Networks may reduce transit dependence. Nearby colocation becomes more useful. Content delivery networks and caches gain additional reasons to locate nearer to users. Enterprises gain additional connectivity choices. Educational and research networks gain new routes for collaboration. Satellite and terrestrial systems gain opportunities for local interconnection. Data centers and compute facilities gain a network anchor.
The value of an IXP, therefore, is not limited to the revenue generated by the exchange itself.
Some of its greatest value may be the ecosystem it enables.
That begins to resemble the broader cooperative and infrastructure-dividend thinking I have been exploring: foundational infrastructure should leave behind durable local capability, not simply consumption.
A Research Opportunity for the Eastern Upper Peninsula
Tagat and Phokeer conclude that more geographically specific research is needed because global data cannot capture all of the local costs, market structures, transport requirements, and institutional conditions surrounding peering.
The Eastern Upper Peninsula presents an interesting opportunity to go further and conduct a deliberately designed rural regional case study before and after IXP implementation.
A useful baseline could include:
- Provider dependence on upstream IP transit.
- Physical and network distance to existing interconnection locations.
- Local versus non-local routing for selected destinations.
- Round-trip latency among regional institutions.
- Number of regional autonomous systems participating in local interconnection.
- Presence of CDNs, caches, and other content infrastructure.
- Number and diversity of interconnection choices available to participating networks.
- Transport costs required to reach existing exchange locations.
- Cross-border routing diversity.
- Percentage of appropriate regional traffic exchanged locally.
- Changes in participating institutions’ interconnection costs over time.
- Growth of nearby data-center, cloud, content, satellite, and network infrastructure.
These measures could be established before development of the exchange and repeated as participation grows.
Rather than asking simply whether an IXP was constructed, the research question becomes:
What happens when a rural border region deliberately develops a neutral interconnection ecosystem as economic-development, resilience, and sovereignty infrastructure?
That question could turn the Eastern Upper Peninsula initiative into something more than an infrastructure project.
It could become a useful contribution to the study of rural Internet architecture itself.
Conclusion
The importance of Dissecting the Cost of Peering to my work is not that it suddenly proves that Internet Exchange Points are beneficial. The technical case for local interconnection is already well established.
Its importance is that it provides empirical evidence for something broader:
Interconnection architecture has economic consequences.
Tagat and Phokeer associate peering savings with broadband pricing, network expansion, and total telecommunications revenue while demonstrating that participation and traffic localization matter. At the same time, their methodological caveats are important. The study should not be treated as proof that constructing an IXP automatically produces particular economic outcomes in every market.
For the Eastern Upper Peninsula, however, the research helps move the conversation from a relatively narrow question:
Why do we need an Internet Exchange Point?
toward a much more consequential one:
Who should have the ability to shape the conditions under which our region participates in the Internet economy?
That is where carrier neutrality, regional economic development, resilience, data sovereignty, tribal self-determination, and community capacity begin to intersect.
An IXP is a place where networks meet.
But a neutral interconnection ecosystem may also become a place where regional agency begins.
Primary Reference
Tagat, Anirudh, and Amreesh Phokeer. Dissecting the Cost of Peering: Evidence from Global Data on Internet Exchange Points (IXPs). Internet Society, Oct. 2026. Paper presented at the 2026 Telecommunications Policy Research Conference.
Related Work
Kronemeyer, Jason. Carrier-Neutral Internet Exchange Point Strategic Implementation Plan: Sault Sainte Marie, Michigan.
Kronemeyer, Jason. International Internet Exchange Grid Tie.
Kronemeyer, Jason. EUPConnect Collaborative Digital Opportunity Analysis Plan.
Internet Society. Navigating the Future: Neutral IXs as Catalysts for Ecosystem Growth.
Provenance Note
The quantitative findings regarding peering costs, IP-transit costs, peering intensity, traffic localization, broadband pricing, network coverage, and telecommunications revenue discussed in this note are derived from Tagat and Phokeer’s Dissecting the Cost of Peering. The authors explicitly identify important limitations involving data availability, peering-cost measurement, comparability of transit and peering prices, geographic coverage, and causal identification.
The carrier-neutral Eastern Upper Peninsula strategy, international-border framing, regional infrastructure relationships, and connections to tribal data sovereignty draw from my previous research, planning, and regional work.
“Interconnection sovereignty,” the framing of an IXP as a regional economic-retention mechanism, the three-layer connectivity/data/interconnection sovereignty model, and the connection to an infrastructure dividend are my synthesis developed in this research note. They should not be attributed to Tagat and Phokeer or the Internet Society without qualification.