RedClara Built a Continental Backbone Through Federation, Not Ownership

Summary

• RedClara is a non-profit association based in Uruguay that federates national research and education networks; it is neither a shareholder company nor an operator that controls each campus.
• Eight main routed nodes and long-term international capacity form a 100G-class backbone, with declared links of 100, 200, and 300 Gbps and a smaller Central American segment.
• ALICE and ALICE2 created the institution and the first network; BELLA secured durable capacity on the direct Portugal–Brazil route, reducing dependence on indirect paths via North America.
• The test for BELLA II is to turn project funding into durable, member-sustained infrastructure, strengthening weaker national networks, public evidence, and supplier diversity.

BELLA changed the network's geography

BELLA, Building the Europe Link to Latin America, combined a submarine component with terrestrial upgrades. BELLA-S secured long-term optical spectrum on a direct cable between Portugal and Brazil, while BELLA-T strengthened the South American backbone. The physical system became EllaLink, approximately 6,000 kilometres between Sines and Fortaleza.

The programme provisioned two 100 Gbit/s links: one between GÉANT and RedClara and another associated with Copernicus data usage. Research and education traffic began using the direct route in 2021. Programme materials mention latency reduction of up to 50% on certain paths; this should be attributed and not generalised to all source–destination pairs.

The BELLA-S1 action registered with the European Union had a cost of €12.4447 million with a European contribution of €10 million. Other figures cover terrestrial components, other directorates, and national contributions, so different totals may be correct in different scopes. Adding them without separating actions would produce a false investment figure.

The 2024 report describes the wider BELLA infrastructure as 35,830 kilometres and Nx100G. This number includes terrestrial and international components beyond the 6,000-kilometre submarine segment. The distinction is necessary so as not to turn a programme measure into cable length.

Forty optical channels do not mean owning the cable

BELLA obtained long-term rights over forty optical channels on EllaLink. This is one of the most important facts in RedClara's history and one of the easiest to exaggerate. Spectrum rights give the community dedicated capacity and upgrade potential, but do not equate to exclusive ownership of the submarine system.

EllaLink is a commercial infrastructure with its own owners, financiers, operators, stations, and customers. RedClara and GÉANT control the contracted rights within that system. They do not single-handedly operate the repair ship, all repeaters, the physical route, every station, or the other services on the cable.

The value of the forty channels lies in optionality. Instead of buying a single fixed capacity, the consortium can light spectrum as equipment and demand evolve, subject to contractual terms and available technology. This reduces the risk of being locked into the initial speed.

The ownership boundary determines responsibility. A fault may require action by the cable operator, station owner, optical supplier, RedClara, GÉANT, or an NREN. Resilience depends on contracts, escalation, and alternative routes, not just the word “direct” on a map.

BELLA and the direct route across the South Atlantic

The most visible change brought by BELLA was geographical. Before the direct route, traffic between Latin America and Europe could pass through North America even when that increased distance and dependence. The Sines–Fortaleza link created a shorter transatlantic alternative for many pairs and a long-term asset for research communities.

The advantage is not a universal promise of halving latency. The gain varies by source, destination, national access, congestion, and return path. The claim of up to 50% reduction applies to specific routes documented by the programme. The more robust result is diversity: a major Europe–Latin America relationship no longer relies exclusively on North American gateways.

This diversity benefits large-volume transfers, astronomy, Earth observation, distributed computing, academic telemedicine, and interactive collaboration. It also amplifies the value of GÉANT and RedClara as interconnection platforms for projects requiring predictability, not just public internet access.

The direct route still depends on physical infrastructure subject to cable, power, station, equipment, and maintenance failures. Resilience requires alternative paths, failover testing, and sufficient capacity in Miami or other routes when EllaLink is unavailable. A direct connection improves the architecture but does not eliminate the need for redundancy.

BELLA also cemented a long-term institutional relationship between Europe and Latin America. Capacity was acquired within a scientific cooperation programme, not as a simple transit purchase. This gives the community greater influence over upgrades and usage, although commercial providers remain essential.

Spectrum rights, terrestrial reach, and the ownership limit

The phrase “RedClara network” hides different forms of control. The organisation may own routers, lease rack space and power, contract an operator, hold an IRU, use spectrum on a third-party cable, or receive capacity through exchange. Each form has distinct duration, risk, upgrade rights, and maintenance obligations.

At sea, BELLA secured long-term spectrum on EllaLink. On land, BELLA-T and the members combined fibre, lambdas, and national services to reach points of presence and institutions. The 35,830 kilometres figure cited in the report is a programme view, not a declaration of full RedClara ownership.

This distinction prevents two errors. The first is minimising the achievement: spectrum rights can be a strategic asset comparable to owning dedicated capacity for many years. The second is exaggerating it: RedClara does not unilaterally control trenching, stations, cable repairs, all equipment, or every national contract.

The true robustness must be read in the contracts. How long does the right last? Who pays for maintenance? Can capacity be upgraded? What happens if the operator is sold, becomes insolvent, or changes strategy? Is there access to stations and equipment? Are there exit routes? These questions determine the real long-term value.

The federated model makes it possible to build reach without tying up capital to own all infrastructure. In return, it increases the need for supplier management and transparency. The organisation must know which dependencies are replaceable and which have become single points that are hard to unwind.

A continental network that is not a single network

RedClara is easily misunderstood when called merely the “Latin American research network.” The phrase suggests a single operator with control from the international backbone to a university socket. The actual architecture is federated. Typically, a researcher reaches RedClara through an institutional network and through a national research and education network, or NREN. RedClara provides the regional and intercontinental layer that allows separately managed systems to work within a larger scientific environment.

RNP, REUNA, RENATA, CEDIA, CUDI, RedCONARE, and RAU maintain their own national governance, domestic infrastructure, institutional relationships, and operational responsibilities. RedClara does not replace them. It interconnects them, represents common requirements, coordinates capacity and services, and creates paths to peer networks such as GÉANT, Internet2, and CANARIE.

An end-to-end communication may traverse the campus LAN, institutional firewall, national access, NREN backbone, RedClara, an international network, the destination NREN, and the remote laboratory. A regional 100 Gbit/s segment does not fix a congested campus edge, misconfigured storage, or restrictive policy at the destination. Likewise, an isolated NREN cannot single-handedly negotiate all routes, identity relationships, and collaborative services required by modern science.

The more accurate description is “shared coordination and infrastructure layer.” Shared means governance, engineering, long-term capacity rights, and common services; it does not mean ownership of every fibre, router, PoP, or cable segment shown on the map. That is how the organisation combines national systems and commercial contracts into a regional network of public interest.

Why Latin America needed a regional layer

The founding problem was fragmentation. In the early 2000s, several countries had national or emerging academic networks, but not a lasting continental organisation capable of aggregating demand, operating a regional backbone, and negotiating on comparable terms with European and North American networks. International circuits were expensive, domestic markets varied, and commercial routes served operators’ economics first, not distributed science.

A university could have good internal connectivity and yet depend on expensive, indirect international transit to collaborate with another Latin American institution. Traffic between Europe and Latin America often passed through North America. The cost was not just latency: there was weaker negotiating power, little route diversity, and reliance on services defined outside the scientific community.

A centralised continental operator would be impractical given the differences in NREN maturity, public funding, regulation, fibre availability, university systems, and landing points. The feasible model was a federation: national networks would preserve legal and operational autonomy, while the regional association would coordinate the common layer and engage with global peers.

The federation also added institutional value. A lone NREN would struggle to influence an international programme or justify a specialised route. A regional organisation could assemble use cases, pool contributions, combine grants with national assets, and present a coherent counterpart to the European Commission, GÉANT, development banks, operators, and cable owners.

Toledo, Valle de Bravo, and the legal association

The institutional idea took shape in 2002 when representatives of Latin American NRENs discussed the viability of a regional network at a CAESAR project meeting in Toledo, Spain. Work continued at gatherings in Rio de Janeiro, Buenos Aires, Santiago, and Mexico, addressing not only topology but also participation, the relationship with national networks, decision-making, and sustainability after external funding.

The legal chronology contains three dates that should remain distinct. RedClara’s history records the signing of the statutes in Valle de Bravo, Mexico, on 9 June 2003. A civil society registration by the Organization of American States uses 10 June as the date of establishment. RedClara states that legal existence was recognised under Uruguayan law on 23 December 2003. These are, respectively, the constituent agreement, an administrative registration, and a later legal recognition.

The most sustained formulation is an international non-profit civil association recognised by Uruguayan law and based in Montevideo. The English institutional phrase “international law organisation” does not prove creation by treaty or the personality of an intergovernmental organisation. RedClara is a membership association with regional objectives, not a supranational authority.

The current statutes were approved in November 2019 and published in January 2020. They define the General Assembly as the supreme body, a Board of Directors, a Fiscal Committee, a Technical Commission with experts from the NRENs, and an Executive Directorate for implementation. Legitimacy derives from membership and cooperation, not from share capital or territorial sovereignty.

Seven members, eleven connected countries, and a wider mission

The current public directory lists seven member NRENs: RNP, RENATA, RedCONARE, REUNA, CEDIA, CUDI, and RAU. The advanced connectivity page, however, reports eleven connected countries: Argentina, Brazil, Chile, Colombia, Costa Rica, Ecuador, El Salvador, Guatemala, Mexico, Paraguay, and Uruguay. Older materials mention thirteen. These numbers measure different things.

Member status defines governance rights, obligations, and participation in the association. Operational connectivity may exist through partnership, transition, or project without the national network being a full voting member. Historical counts may reflect earlier compositions. The correct description separates seven legal members from eleven currently connected countries.

This difference creates strategic tension. RedClara aims to serve a region larger than the formally governing association. A country may receive a connection or participate in BELLA II without equivalent voice in the Assembly; the seven members, in turn, bear a larger share of governance, technical work, and possibly recurring funding. The gap affects legitimacy, cost-sharing, and building a truly regional platform.

For 2025–2027, José Palacios of REUNA chairs the Board; Eduardo Grizendi of RNP is vice-chair; Juan Pablo Carvallo of CEDIA is treasurer; Carlos Gamboa of RedCONARE is secretary; and Eduardo Grampín of RAU is a director. Luis Eliécer Cadenas is the Executive Director. Public documentation does not include complete recent minutes, voting weights, current contribution amounts, or the decision-making role of connected non-member countries.

ALICE provided the regional network's seed capital

The legal formation of RedClara and its first operational network were inseparable from the ALICE project, the Europe–Latin America interconnection. The European Commission and DANTE signed the contract in June 2003. The budget was €12.5 million: €10 million from the Commission and €2.5 million from Latin American partners. The project ran until March 2008.

ALICE simultaneously solved seed capital, common procurement, an operational purpose for the new association, and the link to GÉANT. Instead of waiting for a mature membership and revenue base, it allowed the network’s value to be proven while the institution consolidated.

The first infrastructure was modest by today’s standards, but it created PoPs, production interconnection among NRENs, and relationships with European and North American networks. These relationships later expanded to identity, middleware, grid computing, open science, and joint projects. The operating network turned the association into something more than a forum.

ALICE also created a dependency that would mark its history: a grant can fund construction but not guarantee operation after it ends. Survival would require regional contributions, lower costs, longer-term capacity rights, and an organisation capable of continuous negotiation.

ALICE2 shifted coordination to Latin America

ALICE2 began in December 2008 under an €18 million contract from the European Commission. The decisive institutional change was RedClara taking on the coordination, not merely participating as a regional beneficiary. This shifted implementation leadership, consortium management, and responsibility to the Latin American organisation itself.

The programme sought to strengthen sustainability, broaden participation, reduce recurring costs through dark fibre, wavelengths, and irrevocable rights of use, develop communities, and train technical teams. The move from short-term circuits to long-duration assets was an economic strategy, not just a technological upgrade.

Sources differ on the closure, using March 2012 or January 2013. The safe formulation acknowledges operational completion in the 2012–2013 period and the current project page for later administrative closure. The discrepancy is small but illustrates why announcement, operational, and administrative dates should not be merged.

ALICE2 consolidated RedClara as an institution capable of maintaining a production network after a grant. It also revealed that sustainability is not complete self-sufficiency: even with regional leadership, expansion remained dependent on European cooperation, national assets, and commercial contracts.

From recurring circuits to long-term capacity rights

A regional network’s economics depend on how capacity is acquired. A short-term managed circuit offers a defined service but exposes the operator to renewal, price adjustments, and the carrier’s priorities. Dark fibre, optical spectrum, IRUs, or capacity exchanges require more planning and equipment, but can offer long-term control and a lower marginal cost of upgrade.

RedClara has evolved to combine several forms. It may own routers and equipment, lease space and power, contract maintenance, obtain optical rights on cables it does not own, and exchange capacity with partners. The logical map is therefore a composition of technical control, contractual rights, and third-party infrastructure.

Long-term rights do not eliminate risk. They depend on the cable’s economic life, access to stations, terminal equipment, marine or terrestrial maintenance, partner solvency, and succession clauses. An optical right without funding to light, upgrade, and repair the route can become stranded capacity.

The advantage is aligning the asset’s horizon with the institution’s. Scientific projects and NREN relationships last decades; buying only annual connectivity leaves the entire infrastructure exposed to volatility. The long-term strategy turns part of the grant into a right that can outlive the programme, provided operational obligations remain funded.

Before BELLA, reaching Europe often meant passing first through North America

The lack of a direct high-capacity route between Europe and Latin America shaped performance and dependence. Even with institutions on opposite sides of the Atlantic, traffic could go up to a North American point and then cross the North Atlantic, because that was the commercial fabric available for purchase.

The detour added latency, concentrated failures, and placed the European relationship inside a topology defined by North American hubs. For interactive science, Earth observation, large data transfer, and remote instruments, distance and variability matter as much as nominal speed.

Indirect routes were not technically “wrong”; they were a result of the economics of cables and interconnections. The answer required investment in submarine and terrestrial capacity, not just routing preference. The academic community needed to make its demand large and predictable enough to justify direct rights.

BELLA was conceived to change this geography. The goal was not just more bandwidth, but physical and institutional diversity: a direct link that the research community could use and upgrade over the long term.

The current topology: eight nodes and different capacity types

RedClara’s current technical description names eight main routed nodes: Buenos Aires, Fortaleza, Porto Alegre, São Paulo, Santiago, Manta, Panama City, and Miami. The South American nodes form a mesh of 100 Gbit/s or more; the shared Central American infrastructure is described as 20 Gbit/s. Fortaleza is the main direct path to GÉANT and, through it, to European and African partners. Miami provides the main exit to Internet2, CANARIE, and other North American relationships.

The advanced connectivity page also speaks of 100, 200, and 300 Gbit/s links, eleven countries, and more than 1,300 institutions. On the same site there is still old 10 Gbit/s backbone language. The more recent evidence supports the 100G-class description; the 10G text should be treated as outdated content, not as proof that the modern network does not operate.

The numbers do not mean identical capacity on every path. An optical link may have a different line speed from the protected, reserved, or effectively available capacity for a member. The logical topology may also use multiple providers and shared physical segments. A logical “ring” does not guarantee that all routes are physically diverse.

The presence in Panama and Miami broadens the Caribbean and Central American dimension, but coverage remains dependent on national networks and agreements. Manta provides a relevant point on the Pacific side, while Fortaleza anchors the South Atlantic. The architecture is a result of geography, cable stations, telecommunications markets, and NREN relationships, not a purely symmetrical design.

More than 1,300 institutions is a measure of institutional reach, not an audited census of equivalent connections. The public figure does not explain how many are active, at what speed, or which regional services they use. The system’s maturity should be assessed by topology, capacity, traffic, availability, and usage, not by a single aggregate number.

AS27750, peering, and separation of traffic classes

AS27750 is RedClara’s public autonomous system identity. Through it, the organisation announces or exchanges routes to member networks, research and education peers, services, and international destinations. Public databases show relationships with GÉANT, ESnet, Internet2, CANARIE, RNP, REUNA, RENATA, RedCONARE, CEDIA, and TENET, but these sources are snapshots, not a complete contractual inventory.

The academic connectivity policy is not simply “faster internet.” Eligible traffic may follow private or dedicated paths to NRENs and scientific centres, avoiding general commercial transit where appropriate interconnection exists. The separation reduces costs, improves predictability, and preserves capacity for research applications.

RedClara also offers separation of commercial internet, peering, and global access. This recognises that a university must reach both academic resources and the public internet. The two flows can use different policies, providers, and cost models. A research network is not physically isolated from the rest of the internet; its distinction is created by eligibility, routing, capacity, and governance.

Selective peering enables direct connection where mutual benefit exists, but does not guarantee acceptance of any network. Presence at exchange points or colocation facilities eases interconnection, yet still requires ports, cross-connects, filters, policies, and support. Quality depends on continuous engineering, not just registering the ASN.

The public publication does not offer a complete, current view of RPKI, ROAs, IRR filtering, leak protection, BGP communities, per-peer availability, or incident methodology. For a network that coordinates critical infrastructure, greater transparency about routing security would allow better assessment of resilience without revealing sensitive operational details.

Layer-2 circuits, operations, and the limits of regional responsibility

RedClara offers dedicated layer-2 virtual circuits. A VLAN can make laboratories, data centres, or instruments from different countries appear connected to the same Ethernet network, which is useful for data transfer, distributed research, demonstrations, and requirements that do not fit well with ordinary IP routing.

A layer-2 service crosses several boundaries. The institution provides the local port; the NREN carries the traffic to the regional network; RedClara coordinates the international segment; and the destination repeats the process. VLANs, MTU, redundancy, bandwidth, addressing, security, and change windows must be compatible across all domains.

This chain makes operational responsibility distributed. A problem may lie in campus equipment, the NREN, RedClara, a transport provider, or the destination organisation. A regional NOC needs clear escalation, 24-hour contacts, circuit identification, and demarcation. Without this, the user sees only “the network” and does not know who can restore service.

The organisation identifies engineering, NOC, and security groups, but does not publish a complete SLA methodology, historical repair times, or an independent incident baseline. Old pages mention 100% average availability and 7.7 PB of annual traffic without scope or calculation. These numbers may be signals, but should not be treated as an operational audit.

The federated principle remains: RedClara coordinates what crosses the regional layer and works with the NRENs, but cannot guarantee every component. A professional service requires the boundary to be understandable to the user and the federation to function as a support chain, not a blame-shifting chain.

BELLA II: expansion through co-investment, not a finished map

RedClara and the European Commission signed BELLA II on 14 December 2022. The programme was designed for 48 months, with a European contribution of €13 million within a €28 million plan. RedClara was to mobilise at least an additional €15 million from governments, development institutions, companies, banks, and other partners.

The goal is to connect or strengthen Peru, Costa Rica, El Salvador, Guatemala, and Honduras, with possibilities for Mexico, the Caribbean, Paraguay, Uruguay, Bolivia, and Belize. These names represent programme targets and dialogues; they do not prove that each route was in production at the 3 August 2026 cutoff.

The competitive dialogue opened in 2025 sought a fibre system with a minimum fifteen-year life, initial minimum capacity of 20 Gbit/s in Central America and 100 Gbit/s in Peru, plus a path to expansion. The approach allows proposals for existing fibre, new construction, capacity exchange, or hybrid models.

The flexibility can attract co-investment but makes responsibility more complex. It is necessary to define who builds, who owns, who maintains, what capacity is reserved for research, how a country enters, how operations are funded, and what happens after the grant. A fifteen-year right is sustainable only if maintenance, power, equipment, and upgrades are also funded.

BELLA II should be evaluated by commissioned routes, enforceable rights, accepted traffic, and post-project arrangements. Memoranda, dialogues, and announcements are real steps, but do not equate to active service. The discipline between proposal and completion protects the credibility of the expansion.

From a transport network to a regional research platform

RedClara’s first generation showed that NRENs could exchange traffic over a regional backbone. The organisation then sought to answer what becomes possible on top of that backbone: identity, file transfer, collaborative environments, training, HPC access, open science, cybersecurity, and thematic communities.

This evolution does not abandon the network. It increases its usefulness. A high-capacity path has limited impact if the researcher cannot authenticate, locate a supercomputer, move data, find partners, or obtain support for a cross-border workflow. The application and community layers convert capacity into usable scientific infrastructure.

BELLA II formalises this logic through innovation hubs, testbeds, open calls, training, strategic dialogues, and partner and funder mechanisms. The 2024 agri-food Ideathon gathered 125 experts from 28 countries, and the first HPC environments created a training and prototype pathway for using shared resources.

A testbed is not a production service. It typically has a selected cohort, limited capacity, specialised support, and an experimental objective. It reveals barriers before permanent deployment but can create false expectations if a demonstration is presented as universal availability.

The opportunity is to become a federation of services and tests, not just of transport. This brings RedClara closer to researchers but expands obligations around security, privacy, support, and lifecycle. The organisation moves up the stack without owning all the components: NRENs, universities, SCALAC, clouds, and partners continue to operate the underlying resources.

BIO+IA and the discipline of early-adopter evidence

The BIO+IA Early Adopters programme is a clear example of the new platform function. Opened in 2026 under BELLA II, it selected five projects from six countries for training, specialist support, and access to an AI and bioinformatics environment. Topics included breast cancer biology, molecular screening, antimicrobial or mercury resistance, drug repurposing for liver diseases, and translational biomedicine.

At the July 2026 Demo Day, teams reported how the environment helped research, integrate, and interpret biological information. This is significant evidence of a facilitated research flow and that network, tools, and experts can be combined regionally for data-intensive biomedical problems.

The evidence does not demonstrate clinical efficacy. A model that suggests molecules does not prove a treatment; a fast pipeline does not replace the laboratory, validation, regulation, or health-system deployment. The programme served a limited cohort and results are reported by entities and organisers, not by independent clinical trial.

The qualification defines the correct unit of achievement: RedClara showed it can support structured regional experimentation, reduce tool and knowledge barriers, and build relationships among multi-country teams. The next test is continuity.

A demonstration becomes infrastructure when teams can return, new cohorts enter, data have governance, computing and storage are funded, and successful workflows migrate to national or institutional systems. Repeated use, publications, independent assessment, and post-pilot funding will indicate whether BIO+IA becomes a recurring model.

The broader lesson is to measure innovation by the distance between participation and durable capacity. Ideathons count people and testbeds count experiences; infrastructure also needs to count maintained environments, recurring users, governed data, trained operators, and projects that continue after the grant or the event.

Identity as infrastructure: FIEL, eduGAIN, and eduroam

A network path is useless if the researcher cannot prove identity to the destination service. FIEL supports national federations; eduGAIN enables metadata exchange between federations and international access through home-institution credentials; eduroam applies similar logic to network access while mobile.

These systems do not create a single regional password database. In a federated login, the service redirects the user to their home institution or federation. That institution authenticates and returns a signed assertion or selected attributes; the service applies its policy. RedClara strengthens the federation and connects it to the global trust fabric, without normally receiving the user’s password.

The architecture distributes trust and risk. The university retains responsibility for its accounts, and the researcher reuses the identity across repositories, platforms, and tools. But assurance levels vary, metadata can become stale, and a compromised key or provider can affect services beyond one campus.

eduGAIN solves discovery and metadata interchange, not all legal questions. Each service decides which institutions and levels it accepts; each institution must protect keys, manage the account lifecycle, and respond to incidents. Cross-border use demands consent, attribute minimisation, privacy, and a legal basis.

In eduroam, the visited institution provides connectivity and forwards authentication through the RADIUS hierarchy to the home institution, which validates the credential. Security depends on correct device configuration, certificate validation, home identity, and local controls.

Identity is infrastructure continuity: a supercomputer may be reachable at the IP level and yet unavailable if the trust path fails. Coordinating identity and connectivity addresses a gap that a pure transport network would leave out.

MiLab, eNVIO, and the collaborative research layer

MiLab brings together open tools behind federated access: a data environment based on Dataverse, development with GitLab, communication with Mattermost, and notebook resources linked to JupyterHub, Python, and R. The combination supports a team from communication and code through to analysis, curation, and publication.

Distributed teams often assemble these functions across disconnected services, with different accounts, data policies, and support. A regional space reduces fragmentation and creates a common environment for entities from multiple countries. RedClara has documented uses in collaboration with geographically distributed data.

MiLab should not be called a universal regional cloud. It does not automatically offer large-scale HPC, unlimited storage, certification for any sensitive data, or permanent legal preservation. Capacity, service level, and hosting may change and must be compatible with security, data, and performance requirements.

eNVIO addresses another barrier: moving large files within academic infrastructure. Email and consumer services are inadequate for many scientific datasets. The service’s value depends on limits, retention, encryption, support, and integration — metrics not fully published.

RedClara also operates training platforms and an LMS developed with CEDIA. Courses help turn specialist knowledge into capacity, but do not replace permanent teams and production experience.

This layer sums up the federated model: RedClara integrates tools and access; NRENs bring users; institutions provide identity and context; researchers do the work. The service works when responsibilities are clear and fails when the regional platform is expected to compensate for all local gaps in staffing, governance, and support.

Ventanilla Abierta and navigating a fragmented ecosystem

Research infrastructure can be hard to find even when it exists. A scientist may know there is HPC, storage, identity, or a specialised network somewhere in the region, without knowing who operates it, whether the project is eligible, or how to apply. Ventanilla Abierta was created as a guidance gateway for this problem.

The service does not own all the resources it points to. It translates the need into a pathway: connectivity via an NREN, computing via SCALAC or another facility, data and storage, identity federation, an international resource, or a project and funding route. Its value lies in coordination and discoverability.

This makes it navigation infrastructure. Physical capacity can sit idle because institutional interfaces are fragmented. When the researcher must separately get to know multiple NRENs, rules, and teams, a regional option can be abandoned. A guided entry reduces this transaction cost.

Funds and Partners deals with the complementary side: calls, partner institutions, and funding. Connecting a technical demand to a consortium or funder can turn an idea into a project able to use the network.

The limit is authority. A referral does not reserve capacity; a call does not guarantee an award; an introduction does not create a consortium. The service should be evaluated by successful referrals and sustained projects, not just the number of enquiries.

LA Referencia, SCALAC, and the shared knowledge and computing chain

The open-science and HPC relationships show how transport becomes an ecosystem. LA Referencia aggregates metadata and results from national repositories. SCALAC connects advanced computing organisations. They are distinct alliances, not RedClara subsidiaries, although the network and regional coordination enable cross-border work.

The open-science chain starts at the institution. Researchers deposit publications or data under local policies; national systems organise metadata; LA Referencia creates regional discovery and interoperability; links with initiatives such as OpenAIRE extend international reach. RedClara provides connectivity and coordination, not ownership of every repository or output.

Bandwidth is not enough. Metadata, persistent identifiers, licences, preservation, governance, and incentives determine whether an output can be found and reused. A fast network moves the file, but does not make a poorly documented dataset reproducible.

SCALAC faces a similar challenge in computing. HPC resources are distributed across universities and national centres. A regional alliance makes capacity and expertise more visible, supports testbeds and training, and allows projects to take part without each country funding every category of equipment.

An HPC federation is not a single supercomputer. Queues, schedulers, software, storage, policies, authentication, and accounting vary; sensitive data may face cross-border restrictions. The network removes one bottleneck and exposes the others.

The benefit is the option to share. The risk is maintaining episodic access, dependent on projects and personal relationships, instead of a predictable service with published eligibility, capacity, and cost recovery.

Health, Earth observation, and the communities that give the network purpose

RUTE-ALC brings together university communities in telemedicine and digital health. It can support specialist consultation, medical education, image exchange, standards, emergency collaboration, and digital governance. The network provides paths, identity, and tools; it does not become a healthcare provider.

Clinical privacy, device regulation, professional liability, and consent remain with institutions and authorities. A high-capacity circuit is a telemedicine enabler, not a complete clinical service.

The Copernicus Academy in the region creates another link. Earth observation data can support agriculture, disasters, climate, oceans, water, forests, urban planning, and energy. RedClara helps move data and organise training; the bottleneck is often analytical capacity, not the raw imagery.

TICAL brings together higher-education technology leaders. The 2024 report recorded more than 200 entities in Rio de Janeiro. Meetings are not fibre, but they align procurement, security, cloud, identity, and priorities among organisations that later need to interoperate.

These communities set an NREN apart from a discount operator. Recurring relationships among teams can accelerate projects and incidents. However, attendance does not prove implementation; a workshop does not create national capacity; a connection does not demonstrate clinical or environmental outcomes. RedClara’s contribution is to enable and coordinate conditions.

EduLACSeg and the regional security trust network

Academic institutions face similar campaigns in different countries. eduLACSeg offers a trust framework for sharing indicators, context, and experience, allowing a threat seen by one network to be communicated to others before each discovers it in isolation.

In July 2026, threat information originating at CERN was distributed by RedClara to RNP, REUNA, RENATA, RedCONARE, RAU, CUDI, and CEDIA. It is concrete evidence of a regional intelligence flow. The description of “largest academic cybersecurity ecosystem” is a self-characterisation, without an independent comparison category.

Indicators can be incorrect, stale, or contextual; attribution can be uncertain; privacy and confidentiality laws vary; national capabilities are not equal. Distributing an alert does not guarantee a uniform response.

The 2024 report recorded training with LAC4 for 85 entities from nearly 60 organisations in 18 countries, 30% women, and a Security Baseline Bootcamp with RNP and GÉANT during TICAL2024. These are signs of capacity-building, not proof of equivalent implementation.

RedClara must also protect routing, identity metadata, platforms, and testbeds. There is no complete, current publication of RPKI posture, ROA coverage, IRR filtering, incident methodology, and response times. Regional trust requires classification, redistribution, data-protection, attribution, and post-incident learning rules, accompanied by measurable controls.

SUBMERSE and the network as a scientific instrument

SUBMERSE explored how operational telecoms fibre can become a sensor. Distributed acoustic sensing techniques send optical pulses and analyse the backscattered light to perceive tiny disturbances associated with earthquakes, waves, traffic, or other phenomena.

The idea reuses infrastructure that crosses areas where dedicated sensor networks would be expensive. Research networks can connect scientists to cable owners, stations, data, and processing.

Operational fibre, however, does not automatically become an observatory. Access, instrumentation, calibration, synchronisation, storage, processing, and validation are needed. Sensing and traffic must coexist without compromising service, and the data may reveal sensitive activities.

RedClara was a consortium member, not the owner of all the fibre nor the sole originator of the method. The main funded cycle had its final dissemination in 2026. The outcome is a demonstration and a direction for cooperation, not a permanent, already-deployed regional service.

If future phases secure access, standards, funding, and users, sensing could give RedClara a distinctive role in geophysics, oceans, and resilience — value produced by the infrastructure itself, not just by the content it carries.

A partner ecosystem, not a collection of subsidiaries

RedClara relates to NRENs, the European Commission, GÉANT, Internet2, CANARIE, development banks, EllaLink, operators, SCALAC, LA Referencia, RUTE-ALC, LNet/LACChain, universities, and communities. These relationships are not of the same nature.

An NREN may be a member and an operational partner. GÉANT and Internet2 are peers. The European Commission funds programmes. EllaLink provides commercial infrastructure on which capacity rights exist. SCALAC, LA Referencia, and RUTE-ALC have their own governance. A university can be a user, host, beneficiary, or knowledge provider.

Calling everything RedClara would overstate control and erase responsibility. The association’s value is coordinating assets and entities it does not own. Analysis must distinguish membership, funding, capacity, interoperability, implementation, collaboration, and project participation.

The ecosystem reduces dependence on a single owner but can diffuse responsibility. When a failure occurs, the user needs to know whether it lies in the institution, NREN, RedClara, operator, platform, or partner. The federation is stronger when the boundaries are visible.

Finances based on members, projects, and capacity rights

RedClara has no verifiable shareholders, market valuation, or distributable profit. Its statutes allow member contributions and other resources. There are also consultancy and services, projects from the European Commission and development banks, cooperation, partner contributions, co-investment, exchanges, and in-kind contributions.

The best-known public figures are programme budgets, not institutional revenue. ALICE was €12.5 million; ALICE2 €18 million; BELLA had actions with different scopes; BELLA II was planned at €28 million, with €13 million from the Commission and at least €15 million to be mobilised. Adding all of them as RedClara turnover would be incorrect.

Part of the value lies in spectrum rights, IRUs, national fibre, colocation, equipment, and personnel. These assets need to be assessed by term, maintenance, upgrade rights, and restrictions, not just by the nominal project value.

The 2024 annual report describes activity but does not provide complete audited statements, reserves, funder concentration, or supplier exposure. Also, no exact current staff number was found. This gap limits the assessment of sustainability.

The main risk is temporal: projects last years; the backbone, identity, and platforms need to last indefinitely. Long rights help, but operations, support, power, equipment, and personnel require recurring revenue. Sustainability combines contributions, cost recovery, national support, and partners, without confusing a grant with permanent income.

The last mile is national, institutional, and uneven

A continental network can be sophisticated and yet produce uneven benefit. Countries differ in domestic fibre, NREN funding, public support, competition, cable stations, university, and staffing. Within the same country, a central institution may have excellent access while another relies on a fragile link.

RedClara can aggregate demand, negotiate, train teams, and demonstrate value; it cannot replace national investment, remove all regulatory barriers, or operate every access link. The last mile remains with institutions, NRENs, and national providers.

Hence, “connected country” does not mean uniform quality. The 1,300-plus institutions figure measures reach, but does not disclose methodology, speed, availability, or usage. A mature metric would separate listed, active, capacity-connected, and regional-service-using institutions.

BELLA II can extend the frontier and, at the same time, create obligations where local capacity is weak. Sustainable inclusion requires engineering, budget, support, security, identity, and integration. An international link without a strong NREN can become a high-capacity island.

Documentation debt obscures operational maturity

The public record is rich but brings together incompatible metrics: seven members, eleven countries, and old references to thirteen; a modern 100G-class description alongside 10G text; institution, traffic, and availability figures without consistent methodology.

The likely explanation is documentation debt. Pages from different generations remain live. This does not negate the current network but forces the reader to check source, date, and definition. In infrastructure, outdated documentation can be an operational risk, not just an editorial one.

A map should state the date, contracted and usable capacity, protection, operator, commissioning status, and right type. Traffic needs period, direction, and aggregation; availability needs scope and formula; membership should separate governance, connection, and project.

Coherent information improves decisions by members, funders, and users and prevents memoranda from being mistaken for active links. Transparency does not require exposing secrets; it requires defining precisely what is being asserted.

RedClara's achievement is federation at infrastructure scale

The central achievement is not a cable, router, or programme. It is a durable mechanism by which autonomous national networks aggregate demand, operate interconnection, secure long rights, participate in global projects, and build common services without giving up their governance.

ALICE turned the idea into a network. ALICE2 shifted coordination to the region and pursued sustainability. BELLA changed the geography with the direct route and spectrum. BELLA II tries to widen the reach and convert transport into identity, open science, computing, security, and experimentation.

The federation also imposes limits. RedClara does not own every asset, does not control each NREN, does not guarantee the last mile, and does not have unlimited resources. Its power is coordination: strong when members, funders, peers, and suppliers align incentives; weak when responsibility is diffuse or national capacity is limited public evidence.

Future relevance will depend less on a new speed record than on the institution around the capacity: representation, durable contracts, transparent finances, measurable security, service lifecycles, and a clear path from pilot to production. The backbone is the base; the institutional federation is the strategic asset.

Signposts that BELLA II has become durable infrastructure

The first signpost is commissioning. Memoranda, dialogues, and requirements must result in named suppliers, accepted routes, live traffic, failure procedures, and dates. Capacity must separate the initial minimum, line rate, protected aggregate, and usable bandwidth. A fifteen-year horizon is worth something only with sustainable operations, maintenance, and upgrades.

The second is membership. Today there are seven member NRENs and eleven connected countries. One must track whether the target countries become members, partners, or permanent entities, and how each status affects voting, contributions, rights, and obligations. Expansion without a wider recurring base increases the burden and keeps control concentrated.

The third is operational evidence: dated topology and traffic, SLA methodology, incidents, route security, RPKI, and measured diversity between EllaLink, Miami, and alternatives. A cable cut or route leak will test whether the logical diversity is physical and political in practice.

The fourth is service conversion. BIO+IA, HPC, MiLab, identity, eduLACSeg, and open science should be measured by recurring users, production allocations, support, data governance, and post-grant funding. A pilot becomes infrastructure when the user can plan around it.

The last is financial. Current audited accounts, reserves, contribution and supplier concentration, and the ratio between unrestricted and restricted funds would improve credibility. The question is whether RedClara maintains capacity and services between European cycles without turning weaker networks into temporary beneficiaries.

Scenarios supported by the evidence

In the successful-expansion scenario, Peru and Central American NRENs join durable, scalable routes, broadening the regional base. The European route, North American gateways, and Pacific access support larger instruments and computing.

In the service-federation scenario, connectivity becomes less visible because identity, security, data, HPC, and testbeds become everyday products. Value grows, as do obligations around software, privacy, and support.

In the concentrated-governance scenario, membership stays at seven while projects grow. The operation is regional, but representation, costs, and succession depend on a small core.

In the project-cycle scenario, innovations appear through grants and disappear after they end. The backbone continues, but higher layers oscillate. This is the main risk without a recurring-cost model.

Co-investment with operators can create exchanges, shared construction, and long rights, but deepens reliance on a few suppliers and requires control over maintenance, exit, and upgrades.

In the “network as instrument” scenario, SUBMERSE methods extend to more fibres and generate environmental and geophysical services. It is still prospective but could set RedClara apart beyond commercial capacity.

Who controls the system and who bears the risk

Control is distributed. Member NRENs govern the association; the Board and technical groups implement; national networks control domestic access; peers control their own networks; funders shape programmes; operators, cables, colocation, and suppliers control physical dependencies; universities create demand without always having a direct regional vote.

This avoids a single owner but separates decision from consequence. The Board may approve a programme whose local cost falls on the NREN; a funder sets four years and the members inherit decades; an operator may control the only practical route; the user may not know who repairs the service. These interfaces are the main control surface.

Incentives should recognise durable contributions: financial contributions, fibre, personnel, operations, data custody, and security response. Supported countries need a path to national ownership and recurring obligations. Strong members need assurance that solidarity does not become unlimited transfer without accountability.

Contracts must preserve diversity, upgrade rights, audit, exit, and continuity upon a change of supplier control. Long rights reduce price risk but can become stranded if stations, equipment, or maintenance fail. The critical decision is which dependencies will become hard to unwind.

Decisions that cannot wait for the next grant

First, RedClara must decide whether it will be primarily a connectivity federation or a permanent platform operator. MiLab, identity, security, and AI generate value and obligations around data, software, personnel, and liability. Permanent services need an owner, lifecycle funding, response, and wind-down rules.

Second, it must resolve the gap between legal membership and operational reach. Connected countries need status, voice, and a path to membership or a responsible agreement. Informality eases expansion but leaves it ambiguous who governs, pays, and can leave.

Third, financial transparency should be a governance tool. Accounts, reserves, contributions, restricted funds, and supplier concentration determine survival under a funding or cable disruption.

Fourth, a continuity architecture is needed: diversity, incident escalation, backups, identity and platform recovery, cable alternatives, and succession from project services. The network must preserve essential functions when a grant, supplier, path, or institution fails.

The irreversible risk is accumulating services and dependencies faster than the authority, finances, and evidence to sustain them. The opportunity is to turn BELLA rights, NREN governance, and partnerships into a common good that no country could build alone. Leadership must preserve the distribution and make responsibility explicit.

The long-term meaning of a shared regional backbone

RedClara demonstrates infrastructure sovereignty built on cooperation, not isolation. Regional science gains agency because national networks jointly control a regional layer, maintain strategic capacity, and take part in global partnerships on more balanced terms. Commercial infrastructure remains necessary, but mission, routing, and upgrades are not wholly surrendered to the generic internet market.

The model shows how smaller or unequal institutions aggregate demand without giving up autonomy, and how grants can be converted into rights and organisations that outlive the project. It also shows the federation’s limit: coordination cannot forever compensate for weak domestic networks, opaque finances, or uncertain accountability.

The decisive choice is between merely expanding the map and maturing the institution. More countries and links matter, but the durable asset is the system of governance, operations, identity, data, security, and cooperation around them. If that system becomes transparent, resilient, and genuinely regional, RedClara will remain among the most important shared digital infrastructures in Latin America. Otherwise, the backbone will keep carrying traffic while the higher ambitions depend on the next external programme.