InteRed Panama sells local traffic one peering port at a time

Summary

• InteRed Panama's public record supports the thesis that a Panama-local peering port can be worth paying for when the member can exchange real traffic with local access providers, content caches, root and anycast services, and route-server peers instead of hauling that traffic through upstream transit.
• The strongest current evidence is not a price list. It is the operating footprint: InteRed's live IXP Manager member page and IX-F export, PeeringDB and Internet Society Pulse capacity data, public traffic graphs showing tens of gigabits per second of exchange, route-server peering indicators, and public routing records for the InteRed transit ASN.
• The case is still conditional. Public sources do not disclose member fees, cross-connect charges, per-member traffic savings, churn, service credits, switch utilization by port, or before-and-after latency by network. Without those metrics, the evidence supports a serious economic hypothesis rather than a fully priced investment case.

The port decision starts with routes, not sentiment

Imagine a network engineer in Panama City with a router slot available, an access circuit budget under pressure and a familiar default option: send ordinary traffic to an upstream provider and let the global internet sort it out. The upstream already has contracts, billing, support escalation and route visibility. It is not free, but it is administratively simple. A local Internet exchange port at InteRed Panama only becomes rational if it changes the engineering and cost equation by enough to overcome the extra work.

The customer is not merely buying "internet exchange." The practical unit is a connection to a shared Ethernet fabric, a place in the local peering community, route-server participation where appropriate, access to the exchange's traffic and peering tooling, and the operational right to move selected traffic away from paid transit or longer international paths. The value is not the label on the port. The value is the set of reachable networks behind that port and the probability that a packet bound for a Panamanian access customer, a CDN cache, a root-server instance or a regional anycast node can stay close to Panama.

That unit becomes expensive quickly. The visible port is only the first line item. A member may need a data-center presence or transport into the facility, a router interface with enough headroom, optics, cabling, monitoring, BGP configuration, route filtering, maintenance windows, staff time, abuse and incident contacts, and a process for keeping peering sessions healthy. If the member uses a route server, it still has to understand what the route server is accepting and exporting. If it peers bilaterally with larger networks, it must manage policies and expectations. If it gets the configuration wrong, the cost is not just an invoice. It is outage risk, packet loss, customer complaints and time spent troubleshooting something that upstream transit used to hide.

The question, then, is not whether local interconnection is virtuous. The question is whether InteRed Panama has enough useful route density, local traffic, content presence and operational discipline to make that burden pay. On the public record, the answer is a qualified yes for the right member and an unproven case for the marginal one.

The evidence for "yes" is visible. InteRed's public IXP Manager pages list a live member base that includes Panama access networks, regional network-service providers, content delivery networks, root and anycast infrastructure, and the exchange's own route-server and transit identifiers. The IX-F member export captured on July 5, 2026 lists 30 member entries and 43 interfaces, with seven 1G interfaces, thirty-one 10G interfaces and five 100G interfaces, for 817G of interface speed in the export. PeeringDB's exchange record, which is a different self-reported view of the same ecosystem, shows 29 operational connections and 766.5G of summed port speed. Internet Society Pulse, using PeeringDB-derived data for July 2026, reports 25 ASNs, 766G of capacity, 17 of 25 members peering at the route server, and 24 of 25 members with at least one valid RPKI route-origin authorization. InteRed's own public traffic graphs, captured on July 6, 2026 Hong Kong time, showed current in/out traffic of roughly 80.7G/83.1G, day averages near 68.5G/69.5G, month averages near 77.6G/79.0G, and year averages around 61.1G/61.5G, with the year graph showing a much higher outbound maximum.

Those numbers do not prove anyone's net margin. They do prove that InteRed is not an empty civic symbol. Traffic is moving. Major content and access networks appear in the member base. Route-server participation is material. Panama's internet market has a large enough regulated service universe to make a local exchange relevant: ASEP's SATEL public concession table for service 211, the public internet service classification, showed 58 operating commercial concessions, 52 non-operating commercial concessions, four in compliance review and 114 commercial concessions in total when queried for the public telecom concession table. That does not mean 114 viable BGP networks should join InteRed. It does show that Panama's access market is broader than the handful of names a consumer sees on a bill.

The public record also shows what is still missing. No public InteRed fee schedule was located. Public records do not show member-specific transit savings, exact cross-connect charges, port utilization by member, churn, SLA credits, or a clean time series of routes learned through the route servers. The article's conclusion therefore has to be evidence-strength language, not certainty language: the public record suggests that InteRed's local port is economically credible for traffic-rich members, content-facing networks and technically mature smaller operators. The thesis remains unproven for networks whose traffic is mostly outbound to destinations absent from the exchange, whose access circuit cost to the exchange is high, or whose staff cannot afford the operational work.

The company is a non-profit exchange operator with two public network roles

InteRed Panama is best understood as the association behind Panama's national exchange fabric, not as a retail ISP selling household broadband. Its own website describes it as Panama's Internet Exchange Point, a point for IP traffic exchange. The homepage says InteRed was founded on April 20, 1997 with support from Panama's National Secretariat for Science, Technology and Innovation and the OAS RedHUCyT program. Its public "About InteRed" language in later posts describes a private non-profit association with legal personality granted by Resolution No. 175-PJ-67 on April 24, 1997 and official registration on August 28, 2017 under number 295. The same public material says the association's purpose is to administer the internet exchange node in Panama and related activities that benefit members.

The address evidence is consistent across public sources. PeeringDB's organization page lists InteRed Panama, also known as InteRed, at The Century Tower on Via Ricardo J. Alfaro, suite 401-99, floor 4, in Panama. InteRed's public contact page gives the same building and office context. The company's public persona is PanamaIXP on social channels, and the exchange's operating portal runs IXP Manager, the software used by many internet exchanges to publish member, traffic and peering information.

PeeringDB links the organization to two network records and one exchange record. The first is "InteRed Panama Route Servers" with ASN 10391, an open-policy route-server network. The second is "InteRed Panama Transit" with ASN 272037. The exchange itself is "InteRed Panama" in Panama City. The distinction matters. The directory entity in this article is InteRed Panama Transit, but the economic product under analysis is the broader InteRed local interconnection surface: the transit ASN, the route-server ASN and the exchange fabric together are the public clues that show how the association participates in routing and peering.

LACNIC RDAP identifies AS272037 as a direct allocation and shows InteRED Panama as the registrant. It records the resource's registration date as January 4, 2022. RIPEstat's AS overview for AS272037 identifies the holder as InteRED Panama and marks the AS as announced. BGP.Tools, a public routing-data mirror rather than a regulator or company filing, reports AS272037 as active, with Ufinet Panama listed as an upstream and with public peering visibility toward Cloudflare, Gcore and Meta. That is not a proof of internal architecture. It is public routing evidence that InteRed has a visible BGP surface beyond the older route-server ASN.

AS10391 has a different role. PeeringDB describes it as InteRed Panama Route Servers, marks its type as route server, and lists two interconnection facilities: Ufinet Panama - Urbanizacion Obarrio and Panama Digital Gateway. RIPEstat shows AS10391 as not announced in the same way as a transit network. That is consistent with the route-server function: a route server can be central to exchange operations even if it is not a normal transit network carrying retail internet access.

The most important identity point is what InteRed is not. It is not simply another licensed access provider competing for retail subscribers. It is the meeting point where access providers, content networks, root and anycast operators, enterprise networks and regional carriers can exchange traffic. The economics therefore look different from a normal ISP. A retail ISP tries to recover last-mile capital, customer acquisition, billing and support. An exchange operator tries to make the fabric sufficiently reliable, useful and neutral that members voluntarily carry more traffic across it.

What a member buys is optionality, reach and operating discipline

The member's concrete purchase is a port and the operational system around it. A network connects to the exchange fabric, configures addressing and BGP, decides whether to peer through route servers or bilaterally, and monitors traffic. When the exchange works, the member can keep traffic local when both sides are present and willing to exchange routes. When the exchange fails to attract the right peers, the member still pays for a circuit and a router port while most traffic remains on upstream transit.

InteRed's own website frames the exchange around the ability to manage and negotiate locally generated internet traffic. It says a local interconnection node separates local IP traffic from international traffic, reducing operating costs for ISP operators and improving response times and stability by retaining local traffic in Panama. It also emphasizes diversified membership, including not only service concessionaires and ISPs but hosting providers and content generators. That framing is economically sound, but it is still a claim. The public evidence has to show whether the network mix now resembles the claim.

The current public member mix is the best evidence. The IXP Manager member page lists local access and network operators such as Cable and Wireless Panama, Tigo Panama, Ufinet Panama, Liberty Technologies Corp, Metro Telecom, Galaxy Communications, Sistemas Inalambricos, Trans Ocean Network and other operators. It also lists CDN and content-related members such as Gcore, Cloudflare, Meta, Apple and MeteVerseCloud, as well as LACNIC, LACTLD Anycast Cloud, Netnod, PCH AS42, PCH AS3856 and AS112 Reverse DNS. The exact naming varies across IXP Manager, PeeringDB and Pulse, but the broad pattern is stable: a mix of Panama access networks, regional carriers, global content networks and internet-infrastructure services.

That mix matters more than the raw member count. If a small educational platform, hosting provider or enterprise network buys a port and only reaches another small enterprise, the savings may be marginal. If the same port reaches major Panama access customers, common content delivery paths, DNS infrastructure and the route server, the port has a better chance of replacing paid or inefficient paths. The economic unit is therefore not one port in isolation. It is one port multiplied by the probability that the member's traffic has useful counterparties already present.

Route-server participation reduces a coordination cost. Without a route server, each member must negotiate and maintain bilateral sessions with many other members. With a route server, the member can exchange routes with multiple willing participants through fewer BGP sessions, subject to policy and filtering. Internet Society Pulse reports 17 of 25 members peering at the route server in the July 2026 PeeringDB-derived view. PeeringDB's netixlan API view for the exchange, queried in the same research window, showed route-server peer flags on many of the 29 operational connections. InteRed's peering matrix reported zero bilateral peerings and 253 multilateral peerings in the visible matrix, and it explicitly states that the matrix is based on sFlow traffic accounting and route-server BGP peerings.

That last caveat is important. The peering matrix itself says it detects bidirectional TCP flow between routers at InteRed and cannot prove prefixes are actually swapped between routers. In plain terms, the public matrix is a useful sign of activity and route-server dependence, but it is not a full route table audit. A serious member would still need to check route-server policy, prefix limits, filters, communities, IRR and RPKI expectations, and how much of its own traffic would actually move.

The member also buys discipline imposed by the community. InteRed's public story includes MANRS training and route-security work. Internet Society's case study described the exchange's efforts to encourage good routing practices and the role of training in reducing operational incidents. Pulse's July 2026 view says InteRed itself does not participate in the MANRS IXP program, but it also reports 24 of 25 members using RPKI by its route-origin criterion. Those two facts can coexist: a member base can have strong RPKI adoption even if the exchange has not joined a particular voluntary IXP program. For an engineer deciding whether to plug in, the useful takeaway is that public routing-hygiene signals exist, but they should be verified in member onboarding rather than assumed from a badge.

The revenue logic is sparse in public, but the cost logic is clear

No public fee schedule was located on InteRed's website, IXP Manager document store, PeeringDB record or visible posts during this research. That is the largest gap for a business conclusion. A port can save money only if avoided transit, improved performance or customer retention exceeds membership dues, port charges, cross-connects, colocation, transport, router capacity, staff time and risk. Without published fees, the public record can show value drivers but not calculate member payback.

The public evidence does show how revenue probably scales. InteRed is a non-profit association, so the logic is not venture-style margin maximization. It needs enough recurring member revenue, sponsorship, grants, community support or service income to pay for the fabric, staff, facilities, route servers, monitoring, software, support and upgrades. More members make the exchange more useful and can spread fixed costs across a larger base. Larger ports may justify higher fees or at least higher cost recovery, but a fee schedule would be needed to know whether the association prices by port speed, membership class, access type, traffic volume, or some combination.

The cost base is easier to infer from the public footprint. First, there is switching and routing infrastructure. InteRed's 2024 CDNetworks post says the 2021 move to a more robust and connected data center improved switching capacity to more than 4 Tbps and added redundant services. Its 2025 milestone post says InteRed completed connecting three associates using 100G ports, naming major providers and Gcore in the surrounding context. The IX-F export shows five 100G interfaces in the live member inventory. Ports at that speed imply optics, switch capacity, spares, power and a level of maintenance discipline that a hobby exchange would not carry.

Second, there is facility and transport dependence. PeeringDB's route-server record lists facilities at Ufinet Panama - Urbanizacion Obarrio and Panama Digital Gateway. A member that is already in one of those locations can treat an InteRed port as a local cross-connect and configuration project. A member outside those locations must buy transport or a data-center presence. That access cost can dominate the exchange fee for smaller networks. It also creates supplier concentration: the value of the exchange depends partly on the data-center and metro-fiber market around it.

Third, there is route-server and tooling maintenance. IXP Manager makes member information, traffic graphs, peering matrix, looking glass and the IX-F export visible. Those systems need upkeep, security patching, configuration accuracy, member onboarding, and data hygiene. The source discrepancy between the live IX-F export, PeeringDB and Pulse is a good reminder that interconnection data is not a single immutable ledger. The IX-F export lists 30 member entries and 817G of interface speed. PeeringDB's exchange page displays 26 peers, 29 connections, open peers and 766G total capacity in the rendered view. Pulse reports 25 members and 766G as of July 2026. The discrepancy is not fatal; public interconnection sources often differ because of sync timing, filters, disabled records and naming rules. It does mean a buyer should date-stamp the evidence and ask InteRed for the current operational member list before committing capital.

Fourth, there is staff and community cost. A neutral exchange is partly a technical facility and partly a coordination institution. InteRed's own public description emphasizes discussion, consensus-building, research and common member interests. Internet Society's case study describes a period when the exchange had limited momentum and then renewed activity under new leadership. That history matters because a small-country exchange cannot rely on automation alone. It needs members to respond to route problems, coordinate upgrades, implement routing hygiene and make commercial decisions that may not benefit every network equally at the same time.

For InteRed, the cost of selling a local port is therefore the cost of making the port trusted. The switch has to work. The route server has to be predictable. Members need to believe that the organization is neutral. Larger networks need confidence that smaller members will not create routing instability. Smaller networks need enough value to tolerate fees and fiber charges. The association has to manage all of that while not becoming a retail competitor to its members.

A member's payback is an engineering calculation before it is a finance calculation

The right way to test the port is to begin with a traffic matrix, not with the exchange's brochure. A prospective member should take a representative week of flow data, group destinations by network and geography, then ask which of those destinations appear at InteRed through public peering, the route server or likely bilateral sessions. If the member is a content provider serving Panama households, the presence of Cable and Wireless Panama, Tigo Panama, Liberty-related networks, Ufinet Panama, Metro Telecom and smaller access operators matters. If the member is an access network, the presence of Cloudflare, Meta, Gcore, Apple, PCH, Netnod, LACNIC and LACTLD-related services matters. If the member is an enterprise with mostly cloud-office traffic that never touches those networks locally, the public member list may be less compelling.

The calculation has three layers. The first is avoided transit. If traffic that previously crossed a paid upstream path can move to a settlement-free or lower-cost local peer, the member saves on committed-data-rate headroom, burst billing or upgrade timing. This saving is not automatic because many transit contracts are capacity commitments rather than pure usage meters. If the member is already paying for a 10G commit and only moves a small fraction of traffic, the immediate bill may not fall. The economic benefit may instead appear as delayed transit upgrades, better redundancy, or more room during peak periods.

The second layer is performance. A local route that lowers latency, reduces packet loss or removes unnecessary international detours can improve customer experience even when the transit bill does not fall that month. That matters for streaming, education platforms, gaming, financial services, cloud login flows, DNS resolution and customer-support perception. The Internet Society case study's older Edualianza example is useful because it shows a member treating better local reach and resilience as worth a higher monthly outlay. The example should not be generalized mechanically, but it is a reminder that the finance team may see the port as a new cost while the operations team sees it as protection against churn and complaints.

The third layer is operational leverage. A technically mature member can use the exchange to learn its market. Traffic graphs reveal which peers matter. Route-server sessions simplify initial reach. Bilateral sessions can follow once the member sees enough traffic with a particular counterparty. That learning is valuable in a smaller market because commercial internet choices are often made with incomplete visibility. A local port turns some of that uncertainty into measurement. The member can see what moves, what does not move, which routes are stable, which peers are responsive and which traffic still needs transit.

This is also why a cheap port is not automatically good and an expensive port is not automatically bad. Suppose a 1G member has low traffic but reaches a few critical Panama networks with much better latency. The port can be rational as a resilience and quality tool. Suppose a 10G member moves enough content traffic to delay a transit upgrade by a year. The port can be rational even if monthly fees look high. Conversely, suppose a network buys a 10G port but most of its traffic goes to destinations absent from InteRed, while its staff has to maintain filters and respond to route changes. That port may be economically weak even if the headline price is modest.

InteRed's public evidence gives the member enough reason to run this calculation. It does not complete the calculation. The association's strongest commercial move would be to help prospective members do the math without overselling the result: provide current member and route-server visibility, explain access options to the facilities, offer realistic port-sizing guidance, share aggregate route and traffic examples without exposing member secrets, and state the operational responsibilities clearly. In a trust-based exchange market, a buyer who understands the limits of the port is more valuable than a buyer who joins on vague expectations and leaves disappointed.

Traffic evidence says the fabric is economically alive

Traffic is the strongest rebuttal to the idea that InteRed is merely a directory entry. InteRed's public traffic page exposes day, week, month and year graphs through IXP Manager. The values captured during this research showed current traffic around 80.7G in and 83.1G out. The day graph showed maxima of about 178.2G in and 188.2G out. The month graph showed maxima around 239.3G in and 251.2G out and averages around 77.6G in and 79.0G out. The year graph showed averages around 61.1G and 61.5G, and a much higher outbound maximum of 595.5G.

Those figures should be read carefully. IXP traffic graphs can show aggregate bits crossing the fabric, not necessarily retail customer benefit, avoided transit cost, or performance improvement for a specific member. Direction labels can also be unintuitive depending on how the exchange aggregates interfaces. A traffic spike does not reveal whether the traffic was video, software updates, DNS, cache fill, customer browsing, or a temporary measurement anomaly. Still, sustained tens of gigabits per second across day, month and year views show that the fabric has real usage.

The historical trend is also visible in third-party reporting. Internet Society's Panama case study said InteRed's peak traffic was around 4G before the pandemic, exceeded 10G during early pandemic measures, and later regular peaks were around 8G in that article's timeframe. The same story reported a latency improvement after infrastructure changes, with latency moving from the 20-30 ms range to the 5-10 ms range in one period. Those figures are older than the 2026 traffic graphs, but they show the starting point: InteRed's economic problem was not only adding ports. It was making the exchange meaningful enough that traffic and latency changed.

The 2022 Gcore announcement is a concrete example of content changing the value proposition. InteRed said G-Core Labs became the first CDN provider in its node and that the addition moved more than 10G of traffic from international links to the local exchange, with traffic rising from 9G to 12G over a three-month period in the reported context. The numbers are company-published and should not be treated like audited financials, but they align with the economic mechanism. Content presence can make a local port valuable because traffic that would otherwise leave the country can terminate locally.

The 2022 root-server announcement adds a different kind of value. InteRed said that I-root and K-root instances became available to its membership with support from LACNIC, Netnod, RIPE NCC and Cable and Wireless Panama. The post said round-trip response to those root servers from member networks became far lower, averaging less than 5 ms. Again, that is a company-published claim, but it is technically plausible and economically relevant. DNS root-server proximity does not carry the volume of video traffic, but it improves resilience and basic internet responsiveness. For an operator, that can be part of the justification for joining even when raw traffic savings are not the whole case.

The 2025 milestone post is the clearest capacity sign. InteRed said it had connected three associates to the exchange using 100G ports, naming major local providers and Gcore in the narrative. The public IX-F export supports the presence of 100G interfaces, including access and content members. That is important because port speed sends a market signal. A member does not need a 100G exchange port unless it expects significant traffic, resilience value, or strategic positioning. The exchange's job is to turn those large member commitments into value for the smaller participants too.

Member density is still the central risk

InteRed's thesis depends on density. A local port becomes valuable only if enough of the traffic a member cares about is reachable through the fabric. That is why the assignment's phrase "one peering port at a time" is useful: every new port is both a customer and a reason for other customers to join.

The public data shows encouraging density but not saturation. Internet Society Pulse's country page says Panama had two active IXPs with a combined total of 30 members as of July 2026. It also says 68% of Panama's 97 active networks were either IXP members or customers of IXP members, and 14% of the 1,000 most visited websites in Panama were accessible through an in-country server or cache. Pulse lists InteRed with 25 members and BGP.Exchange - Panama City with five. That country-level view supports the view that IXPs matter in Panama, but it also shows room for growth: most popular sites are still not locally cached by Pulse's measure, and a significant share of active networks is not directly at the exchange.

ASEP's public concession data widens the lens further. Service 211, public internet access, had 114 commercial concessions in the SATEL table, with 58 operating. That number should not be mapped one-for-one to BGP-ready exchange prospects. Some concessions may be small, inactive, reseller-like, wireless-focused, geographically distant or operationally unable to justify a port. But the gap between 58 operating internet concessions and roughly 25-30 exchange participants suggests InteRed's market still has a long tail. The economic question is whether the next ten or twenty members add meaningful routes and traffic, or whether they add operational overhead without enough traffic to improve the fabric.

The member mix also creates dependence on a few large names. PeeringDB and the IX-F export show 100G-class participation from large access and content networks. Those members make the exchange valuable, but their presence can concentrate traffic and bargaining power. If a major content member changes caching strategy, reduces public peering, or shifts traffic to private interconnects with large access networks, aggregate exchange value could fall for smaller members. If a major access provider limits what it shares through the route server, a small member may still need transit for the traffic it hoped to localize.

This is where the route-server evidence is most useful and most limited. A high route-server participation ratio lowers coordination cost for new members. But route-server participation does not guarantee that every important route is available, that every prefix is accepted, or that every bilateral commercial preference disappears. A network engineer should ask for current route-server route counts, accepted prefix counts, max-prefix policies, filtering rules, BGP community support, RPKI behavior and member-specific traffic projections. Public pages show the door is open; they do not show the full menu.

Panama's regulation makes the market visible, but it does not sell the port

Telecommunications regulation is relevant because the IXP's potential customer base is made of regulated or at least publicly visible networks. ASEP classifies telecom services in Panama and identifies service 211 as public internet service, defined as a service that allows a customer to connect with the global internet. ASEP's concession process page says Type B telecom services are granted without a public-tender formality to applicants that meet the regulatory requirements, and lists obligations that may include interconnection requirements, the ability to install and operate telecom networks, technical compliance, visible public pricing, reliable measurement systems, state inspection and other concession duties.

That regulatory context has two effects. First, it makes the operating universe more transparent than in a fully informal market. Public concession counts, resolutions and classifications help identify how many providers might need interconnection, transport and routing arrangements. Second, it creates compliance overhead for the same networks. A small operator deciding whether to join InteRed must already manage concession duties, customer service obligations and technical reporting. A local exchange port is not a regulatory requirement in itself. It has to compete for budget and staff time against those obligations.

The local exchange can also support policy goals without being a policy substitute. Keeping local traffic local can improve resilience and reduce dependence on international paths during some disruptions. But an exchange cannot fix last-mile competition, spectrum policy, consumer pricing, power reliability, fiber cuts, or emergency service restrictions. InteRed's public value is strongest when it is framed as practical infrastructure for operators, not as a universal answer to Panama's connectivity policy.

Geography adds another layer. Panama is a regional logistics and connectivity hub, and international capacity into and through the country is part of its broader digital appeal. That can cut both ways. Strong international connectivity can reduce the urgency of local peering if transit is cheap and reliable. It can also make Panama a better place to host caches, anycast services and regional interconnection because international backhaul and local access density meet in the same market. InteRed's 2024 CDNetworks announcement explicitly leaned on Panama's regional position. The public test is whether more such content and edge providers appear in the member base and whether their traffic remains visible on the fabric.

The competitor is often habit, not another exchange

InteRed's most direct substitute is not always another IXP. It is the existing upstream transit contract. Transit is predictable, globally complete and operationally familiar. If a member pays a competitive blended transit price, has limited local traffic, and lacks staff to maintain BGP policy, the IXP port may feel like optional complexity. That is the hurdle every exchange faces.

There are other substitutes. A large access provider can build private interconnects with a major content network and keep that traffic off the public exchange. A CDN can place caches inside the largest access networks rather than peer openly. A local enterprise can buy managed internet and avoid becoming a BGP network. A regional carrier can carry traffic to Miami, Bogota or another interconnection market where more peers are present. A second local exchange, BGP.Exchange - Panama City, appears in Pulse with five members, and its model may appeal to some networks even if InteRed remains the larger visible exchange. Remote peering and cloud on-ramps can also change the traffic decision for members that mostly need access to global cloud platforms rather than local consumers.

InteRed's advantage is that it already has the harder side of the market: major local access networks, major content names, root and anycast infrastructure, and public traffic. An exchange with a prettier price but fewer useful counterparties would not necessarily be cheaper in economic terms. The operating cost of a port must be compared with the traffic it can actually move. A dense port can be expensive and still be rational. A cheap port with no routes can be waste.

For smaller members, the strongest selling point may be risk reduction rather than pure transit savings. Internet Society's case study described Edualianza, a local educational platform, as having joined InteRed after relying on two ISPs and experiencing problems with one provider. The article reported that Edualianza's costs increased because of monthly fee and fiber charges, but the service improved and latency dropped in its measurements. The details are one member's reported experience, not a universal formula. But the example captures a real buyer motive: a local port can add path diversity and a better route to local users, even if the accounting line item rises.

For large access providers, the selling point is different. They may already have private arrangements with content networks and enough scale to negotiate transit. Their reason to participate can be national resilience, lower cache delivery cost, support for local content, route hygiene, or influence over a shared infrastructure they need to remain credible. The 2025 100G upgrades by major local providers are important because they show that large operators still see value in the shared fabric.

The transit ASN is evidence, not the whole business

The directory entity name includes "Transit", and AS272037 is visible as InteRed Panama Transit in PeeringDB. It would be easy to overread that label and treat InteRed as a conventional transit provider. The public evidence does not support that as the main thesis. The stronger evidence points to exchange economics: route servers, member lists, IX-F export, peering matrix, traffic graphs and content/root-service additions.

AS272037 still matters. LACNIC RDAP ties the resource to InteRED Panama. RIPEstat says it is announced. BGP.Tools shows an upstream and a small visible peer set, with a 30G presence at InteRed Panama in its public view. PeeringDB's organization record links that ASN to the same organization as the route-server ASN and exchange. Those records support directory identity and public routing surface. They do not show revenue, customer contracts or internal network architecture.

The distinction is important for business judgment. If InteRed's economics depended mainly on selling transit, it would compete against carriers on price, capacity and global reach. If its economics depend mainly on the exchange, the product is coordination: making a local port worth the work because enough useful networks meet there. The public evidence supports the second interpretation much more strongly.

That interpretation also handles supplier dependence more accurately. Public routing mirrors show AS272037 with an upstream relation to Ufinet Panama. PeeringDB route-server facility data lists Ufinet Panama - Urbanizacion Obarrio and Panama Digital Gateway as facilities. Ufinet also appears as a member with public port capacity. That could mean Ufinet is an important infrastructure partner, facility provider, upstream or all of those in different contexts. Public data cannot safely infer the contract terms. It can only show that Ufinet is visibly present in the exchange's operating surface.

For a member, the practical question is not whether InteRed has a transit ASN. It is whether joining the exchange reduces dependence on the member's own transit providers. If the member's traffic to local access networks, CDNs and DNS services can move over the exchange, then the port can reduce marginal transit usage and path length. If the member mostly sends traffic to destinations absent from the exchange, AS272037's existence does little for the member's cost base.

What would prove the investment case

The public record supports a strong qualitative case, but a true investment decision needs missing numbers. The first missing number is the fee schedule. InteRed should be able to tell a prospective member the monthly and one-time costs for membership, 1G, 10G and 100G ports, cross-connects where applicable, remote transport options, route-server participation and any support tiers. Without this, no member can compare the port against transit dollars per Mbps.

The second missing number is member-specific traffic projection. A member needs to know what share of its current traffic is likely to move to InteRed. That requires NetFlow or sFlow analysis against the current member list, route-server tables and traffic destinations. Aggregate exchange traffic is encouraging, but an operator with mostly international enterprise SaaS traffic may see a lower benefit than a hosting provider serving Panama access customers or a content network delivering to those customers.

The third missing number is latency and path improvement by destination. Public sources show historical latency improvements and root-server response claims, but a buyer needs current measurements to its own customers and peers. A sensible onboarding package would include RIPE Atlas measurements, looking-glass tests, traceroutes, route-server visibility and before-and-after hop counts to major Panama access networks and content endpoints.

The fourth missing number is reliability. Exchange traffic graphs show usage, not outage history. A member should ask about switch redundancy, route-server redundancy, maintenance windows, incident history, power and facility resilience, DDoS handling, escalation process, contact coverage, and what happens when a member leaks routes or violates policy. InteRed's public traffic, peering matrix and route-security narrative are positive signals, but serious buyers need the operational record.

The fifth missing number is churn and member satisfaction. The public member list shows many joined dates, including original 1997 members and new content additions, but it does not show why members leave, downgrade, upgrade or remain inactive. Churn would be a powerful economic signal. If members retain ports and upgrade speeds after measuring value, the case strengthens. If members join for public-policy reasons and then keep traffic low, the case weakens.

Public evidence used in this assessment

The key evidence is visible and mostly current. InteRed's homepage at https://intered.org.pa/ supports the organization's self-description as Panama's IXP, its founding history, local-traffic rationale, open reciprocal exchange language and membership mission. The contact page at https://ayuda.intered.org.pa/contactus supports the public address. PeeringDB's organization page at https://www.peeringdb.com/org/17803 supports the organization alias, address, website, route-server network, transit network and exchange relationship. PeeringDB's exchange page and API at https://www.peeringdb.com/ix/1896 and https://www.peeringdb.com/api/netixlan?ix_id=1896 support the exchange capacity, connection and route-server peer evidence.

InteRed's IXP Manager member page at https://portal.intered.org.pa/customer/details and IX-F export at https://portal.intered.org.pa/api/v4/member-export/ixf/1.0 support the current member list, joined dates and live interface-speed inventory. The public traffic page at https://portal.intered.org.pa/statistics/ixp supports aggregate traffic levels. The peering matrix at https://portal.intered.org.pa/peering-matrix supports the visible multilateral peering count and, importantly, states the measurement caveats.

Internet Society's case study at https://www.internetsociety.org/issues/ixps/success-stories/panama/ supports the historical revival narrative, earlier traffic and latency figures, Edualianza's reported experience, and the importance of CDNs and root servers. Internet Society Pulse pages at https://pulse.internetsociety.org/en/ixp-tracker/ixp/577/ and https://pulse.internetsociety.org/en/ixp-tracker/country/PA/ support the July 2026 PeeringDB-derived view of capacity, membership, route-server participation, RPKI adoption, country network coverage and local cache share.

InteRed's own posts support specific operating milestones: https://intered.org.pa/2022/06/21/g-core-labs-llega-a-panama/ for the Gcore CDN addition and reported traffic impact; https://intered.org.pa/2022/09/16/servidores-raiz-i-y-k-septiembre-2022/ for I-root and K-root deployment to members; https://intered.org.pa/2024/06/03/intered-panama-and-cdnetworks-join-forces/ for CDNetworks and the association's legal description; and https://intered.org.pa/2025/01/06/a-milestone-in-panamas-connectivity/ for 100G member connections and the 27-year history framing.

Public registry and routing evidence includes LACNIC RDAP at https://rdap.lacnic.net/rdap/autnum/272037 for AS272037's registrant and registration event, RIPEstat at https://stat.ripe.net/data/as-overview/data.json?resource=AS272037 and https://stat.ripe.net/data/as-overview/data.json?resource=AS10391 for AS holder and announcement status, and BGP.Tools at https://bgp.tools/as/272037 for a public routing mirror. PCH's IXP page at https://www.pch.net/ixp/details/176 supports status, establishment date, media type and managed-by context. ASEP pages at https://asep.gob.pa/direcciones/telecomunicaciones/servicios/, https://asep.gob.pa/direcciones/telecomunicaciones/solicitudes/ and the SATEL concession table at https://satel.asep.gob.pa/Pages/Control/BucardorConcesionariosTelecomunicaciones/BuscadorConcesionariosTelecomunicaciones.aspx support the regulatory setting, service 211 definition, Type B concession process and concession count.

The main uncertainty is not source authenticity. It is granularity. Public sources are strong enough to identify InteRed, the exchange footprint, member composition, port capacity, aggregate traffic and regulatory context. They are not granular enough to price a port or prove member-level savings.

The judgment

The evidence supports the view that InteRed Panama can make a local peering port worth paying for when the member's traffic intersects with the exchange's visible strengths: Panama access networks, major content networks, root and anycast infrastructure, route-server reach and a fabric carrying sustained aggregate traffic. The public record suggests InteRed is past the point where the pitch depends on national aspiration alone. It has member density, 100G-class commitments, current traffic, route-server activity and enough technical transparency to make a serious buyer run the numbers.

The available evidence is also consistent with a more cautious claim: InteRed's economic value is uneven. The exchange is strongest for networks that serve Panama users, host content, carry traffic for local customers, operate their own routing, or need resilience against single-provider dependence. It is weaker for organizations that lack BGP competence, have little Panama-bound traffic, cannot reach the facilities cheaply, or already obtain adequate private interconnects through a larger provider.

The thesis remains unproven without four named metrics: current public port and membership pricing, per-member traffic shifted from transit to the exchange, before-and-after latency by representative destination, and churn or upgrade history by member class. If those metrics are favorable, InteRed's business case becomes straightforward: a member is paying for a local route market that reduces international dependence and improves service quality. If those metrics are weak, the port becomes a well-intentioned overhead line.

For now, the balance of public evidence favors the thesis, with a condition attached. InteRed Panama does not sell "local traffic" as a commodity sitting on a shelf. It sells the chance to make local traffic local, one peering port, one route-server session and one operationally competent member at a time.