Coalition networks: CENTRIXS, the Mission Partner Envir...

A U.S. commander in a coalition operation faces a problem that sounds mundane and is not: how do you let an allied partner see the shared operational picture, the chat traffic, and the command-and-control data at machine speed, while making sure a different partner who is not party to that agreement never sees it? Foreign-disclosure law does not permit "share everything." It permits sharing specific data with specific partners at specific classification levels. For a quarter-century the U.S. military answered that problem the way engineers under deadline usually do: it built a separate network for each partner group. The result is a family of coalition networks the Pentagon has now spent years and several consolidation programs trying to merge back together.

This is a documentary look at that effort: the Combined Enterprise Regional Information Exchange System (CENTRIXS), the Mission Partner Environment (MPE) meant to consolidate it, and the money that public records actually establish. The recurring trap in this topic is money that is not what it looks like. The most widely cited figure, a $998 million contract, is a spending ceiling, not spending. The most widely cited billion-dollar number belongs to a broader effort and is a budget request, not an obligation. Both matter, and neither is "the cost of CENTRIXS." Keeping them straight is most of the honest accounting here.

What CENTRIXS is, and why it fragmented

CENTRIXS is not one network. It is a family of separate classified coalition networks, each called an enclave, and each operating at a single classification and releasability level. Within an enclave, U.S. forces and a defined set of partner nations get email, web, chat, a Common Operational Picture, and voice. Between enclaves, by design, there is often no interconnection at all. Development began around 1999, and a dedicated DoD program office was established in January 2002 to support the combatant commands, according to Navy and DTIC sources summarized publicly. Those dates are well attested but come from secondary compilations.

The enclave names map to partner groups. Widely documented ones include CENTRIXS Four Eyes (CFE), covering the United States, United Kingdom, Canada, and Australia; CENTRIXS-J with Japan; CENTRIXS-K with South Korea; CENTRIXS-ISAF for the Afghanistan coalition; CENTRIXS-GCTF for the Global Counter-Terrorism Task Force; and maritime enclaves CENTRIXS-CMFC and CENTRIXS-CMFP for combined and cooperative maritime forces. Each is typically a physically separate network with its own terminals. Enclave rosters and exact partner memberships shift with operations, and some are sensitive, so the load-bearing point is not any particular roster. It is the design principle: controlled releasability to specific partner groups, so that data reaches exactly the nations authorized to see it and no others.

That design principle is the honest reason the sprawl exists. Each enclave was stood up precisely to prevent inadvertent release to a nation outside a given agreement. The stovepiping was partly deliberate, and it was lawful.

The scale, and the sprawl

Deliberate or not, the family grew large. DISA's own Multinational Information Sharing (MNIS) briefing describes the CENTRIXS family as spanning more than 40 networks or communities of interest, with at least 80 partner countries plus NATO participating across the enclaves, and connections at more than 40 control sites. These are round program characterizations rather than audited counts, and DISA does not publish a precise current total, so the correct framing is "on the order of 40-plus networks and 80-plus partners," not an exact figure. It is worth noting that the "40-plus networks" count and the "40-plus control sites" count are two different measurements that happen to land near the same number; they are not the same thing.

What that scale means for an operator is captured in one quote that has become the emblem of the problem. Lt. Col. David Courter of DISA, the agency's chief of combatant command plan integration, put it plainly to DefenseScoop in June 2026: "We cannot continue to have 11 different SIPR networks on the island of Oahu and five different versions of CENTRIX[S]." That is a snapshot of one location, not a DoD-wide audited metric, but it describes the lived cost of the architecture: multiple terminals per desk, duplicated hardware, networks that do not interconnect, and slow partner onboarding.

The governance is as fragmented as the wiring. Consolidation has to satisfy the releasability rules of many partners at once. The NATO-led Federated Mission Networking effort, a distinct federation often discussed alongside coalition networking, spans roughly 40 nations and affiliates, as AFCEA reports the figure. FMN is not CENTRIXS, but its membership size illustrates why merging coalition networks is a political and legal problem as much as a technical one.

The fix on paper: MPE, MPE-IS, and the Coalition Information Environment

The Mission Partner Environment is the consolidation initiative. Its goal, and that of the Mission Partner Environment Information System (MPE-IS), is to fold the coalition-networking portfolio into a single, scalable, data-centric environment that preserves releasability controls while retiring dozens of physically separate networks. The policy backbone is DoD Instruction 8110.01, "Mission Partner Environment Information Sharing," reissued on June 30, 2021 (an older 2014 version still circulates online, so the current date matters when citing it).

Ownership is where this gets complicated, and where eras diverge. The legacy DISA MNIS framing was three capabilities: CENTRIXS, Griffin, and CFBLNet. The newer consolidation era is broader and differently owned. The Air Force Mission Partner Capabilities Office (MPCO) is responsible for modernizing, integrating, and consolidating a wider portfolio that includes CENTRIXS, CFBLNet, BICES and BICES-X, Pegasus, APAN, and MPE-IS. Governance across the whole thing is split among the Joint Staff, DISA, the Air Force MPCO, and the services. That split is itself part of the difficulty.

The infrastructure meant to deliver MPE is DISA's Coalition Information Environment (CIE). As reported by DefenseScoop in June 2026, DISA is deploying CIE starting in U.S. Indo-Pacific Command and scaling to European Command, with a demonstration planned during the Olympus Fires exercise in Fall 2026, a minimum viable capability release targeted for January 2027, and a Global Network Operations and Security Center approaching initial operational capability. Every one of those milestones is a plan or target as of mid-2026. Consolidation programs in this space have a long history of slipping, so "planned" and "targeted" are the operative words.

The money, kept distinct

Two dollar figures dominate coverage of this topic, and they measure different things in different years. Neither is a total cost of CENTRIXS or MPE, because no single audited figure of that kind exists in the public record.

The $998 million MPE-S contract ceiling. DISA's requirement for Mission Partner Environment Services (MPE-S) was placed on a single-award IDIQ contract with Trace Systems Inc. carrying a ceiling of $998 million, to design, implement, and operate the enterprise for DoD and its mission partners, supporting on the order of 45,000 users. Two framing corrections matter. First, an IDIQ ceiling is a maximum the government may order against, not money obligated or spent; actual spending lives in the individual task orders beneath it. Second, per the contractor's own description, this vehicle was executed by the Air Force Mission Partner Capabilities Office with a period of performance running from September 2019 to September 2024. It is therefore a past ceiling on an expired vehicle, not current DISA spending. The current infrastructure push is the separate CIE effort with the January 2027 target.

The $1.4 billion CJADC2 budget request. In GAO-25-106454 (2025), the Government Accountability Office reports that DoD requested more than $1.4 billion for Combined Joint All-Domain Command and Control (CJADC2) activities in its fiscal year 2025 budget request. GAO lists U.S. European Command's Mission Partner Environment among combatant-command data-sharing systems, which is how MPE appears in that report. The $1.4 billion is a request, not obligated spending, and it covers the wider CJADC2 effort, not MPE or CENTRIXS specifically. GAO does not isolate a single MPE or CENTRIXS cost. (A separate task-order ceiling near $1.4 billion for CJADC2 modernization has surfaced in 2026 reporting; that is unrelated money from a different mechanism and year, and should not be confused with GAO's FY2025 request figure.)

The two headline numbers, in short: $998 million is a past IDIQ ceiling, not spent; $1.4 billion is an FY2025 request for a broader effort, not spent and not MPE-specific. Merging them into one "cost of the program" would misstate the record in exactly the way the record invites.

What GAO actually found

The oversight thread is where the primary sources are strongest. GAO-25-106454 (2025) does more than list a budget number. It finds that departments and commands "continue to develop and acquire capabilities... largely in isolation," and it recommends DoD establish a comprehensive framework to guide the effort. MPE sits inside that finding as one of several data-sharing systems that grew up without a unifying architecture.

This is not a new pattern for DoD information environments. Nearly a decade earlier, GAO-16-593 (2016) found that DoD needed to strengthen governance and management of the Joint Information Environment (JIE), measuring it against leading practices for scope, cost, schedule, workforce, and security planning. JIE is a related but distinct initiative from MPE, so GAO-16-593 is precedent, not a direct MPE finding. Read together, the two reports describe the same recurring class of concern across different programs and years: governance and cost-and-schedule discipline lagging behind ambition.

The regional stepping-stones: CPN and CPN-X

Before the department-wide MPE push, consolidation happened regionally. In U.S. Central Command, coalition enclaves such as CENTRIXS-ISAF and CENTRIXS-GCTF were integrated into the CENTCOM Partner Network (CPN), a regional roll-up of coalition networks in the CENTCOM area of responsibility. That specific ISAF-and-GCTF-into-CPN integration is sourced to a specialist secondary blog citing an army.mil article that could not be independently re-read, so treat the exact mapping as medium confidence; the CPN consolidation in aggregate is better attested.

A related program name, CPN-X, appears in a primary DoD document. A U.S. Air Force fiscal year 2021 budget justification lists "mission partner networks such as but not limited to: CENTCOM Partner Network (CPN), CENTCOM Partner Network Extended (CPN-X), Mission Partner Environment Information System (MPE-IS)," describing them as secure networks for bilateral and multilateral information sharing up to the SECRET RELEASABLE level, with AFCENT funding its portion to ensure interoperability. So CPN-X resolves to CENTCOM Partner Network Extended, a CENTCOM-area coalition-network extension. Two hedges belong on that identification: it was read via a third-party mirror of the budget document rather than the original comptroller PDF, and no separate dollar figure is broken out for CPN-X. One disambiguation matters too. CPN-X here is a coalition network; it is not the Army WIN-T "Command Post Node (CPN)," an unrelated tactical transport node that shares the abbreviation. Some secondary write-ups also mis-expand CPN-X as "Coalition Partner Network"; the budget document's own wording is "CENTCOM Partner Network Extended," and that is the phrasing to trust.

The honest critique and the honest defense, side by side

The efficiency critique. The coalition-networking mission has been rebuilt repeatedly for a quarter-century. CENTRIXS grew ad hoc from 1999 as combatant commands stood up separate enclaves per coalition and per operation, producing more than 40 stovepiped networks that cannot interconnect by design, leaving operators with multiple terminals per desk and slow partner onboarding. Courter's snapshot of Oahu, 11 SIPR networks and five versions of CENTRIXS at one location, is the emblem. The Pentagon then layered consolidation on consolidation: regional roll-ups like CPN and CPN-X, then the Joint Information Environment that GAO-16-593 found lacked governance and cost-and-schedule discipline, and now MPE, MPE-IS, and the Coalition Information Environment, with a past $998 million services ceiling behind them and a minimum viable capability still targeted for January 2027. GAO-25-106454 reports the wider CJADC2 effort, more than $1.4 billion requested in FY2025 alone, still has commands building "largely in isolation" without a comprehensive framework. The critique is not that any one system failed. It is that decades and large, hard-to-total sums went into building a tangle and then repeatedly paying to untangle it, with duplicative terminals and interoperability gaps persisting through multiple unification initiatives.

The mission defense. Coalition information sharing is a genuine warfighting necessity, not bureaucratic excess. The United States rarely fights alone, and fighting alongside allies requires sharing a common operational picture, intelligence, and command-and-control at machine speed. The reason it is hard, and the reason it fragmented into enclaves in the first place, is legitimate. Foreign-disclosure law and releasability rules require that data be shared only with the specific partner groups authorized to see it, at controlled classification levels. "Share everything" is not the goal; it would be unlawful and dangerous. Each enclave existed precisely to prevent inadvertent release to nations outside a given agreement, and that segmentation kept coalition operations lawful across Iraq, Afghanistan, and maritime counter-terrorism while letting more than 80 partner nations plug in. MPE, MPE-IS, and the Coalition Information Environment are the deliberate, engineered answer: keep the releasability controls, but deliver them through one scalable, data-centric environment instead of dozens of physically separate networks, reducing operator burden and speeding partner onboarding while still honoring disclosure rules. The difficulty is structural, born of alliance politics and law, not merely of mismanagement.

Both verdicts are true at once. The record shows real duplication, real governance gaps that GAO has flagged twice a decade apart, and real money spread across ceilings and requests that resist a clean total. It also shows a mission that is legally required, technically hard for legitimate reasons, and now being addressed by a consolidation program that has not yet had to prove it will land on schedule. The honest test is not whether coalition networking should exist. It is whether MPE and CIE finally deliver the unified framework GAO says has been missing, or become the next layer someone has to untangle.

Fact-check notes and sources

CENTRIXS is a family of separate classified coalition enclaves, each at a single classification and releasability level, providing email, web, chat, a Common Operational Picture, and voice; development began around 1999 with a program office established January 2002. DISA's MNIS briefing and DTIC materials describe the enclave design; history is corroborated via Wikipedia's CENTRIXS summary and Navy sources.
Named enclaves (CFE for US/UK/Canada/Australia, CENTRIXS-J, CENTRIXS-K, CENTRIXS-ISAF, CENTRIXS-GCTF, maritime CMFC/CMFP), each typically a physically separate network. Wikipedia: CENTRIXS; U.S. Navy CHIPS on CENTRIXS-Maritime.
The CENTRIXS family spans more than 40 networks or communities of interest, at least 80 partner countries plus NATO, and more than 40 control sites, per DISA's own MNIS overview. These are round program characterizations, not audited counts; the network count and control-site count are distinct measures. DISA MNIS briefing (Cotter).
Lt. Col. David Courter (DISA): "We cannot continue to have 11 different SIPR networks on the island of Oahu and five different versions of CENTRIX[S]." One location, illustrative of sprawl, not a DoD-wide metric. DefenseScoop, June 3, 2026.
The Federated Mission Networking effort spans roughly 40 nations and affiliates, as AFCEA reports; FMN is a NATO-led federation distinct from CENTRIXS. AFCEA SIGNAL.
MPE and MPE-IS are the consolidation initiative; the Air Force Mission Partner Capabilities Office manages a portfolio including CENTRIXS, CFBLNet, BICES/BICES-X, Pegasus, APAN, and MPE-IS. Policy backbone is DoD Instruction 8110.01, reissued June 30, 2021. DoDI 8110.01 (PDF); MPE lexicon (DoD CIO).
DISA's Coalition Information Environment underpins MPE, deploying first in INDOPACOM and scaling to EUCOM, with a Fall 2026 Olympus Fires demonstration, a January 2027 minimum viable capability target, and a GNOSC nearing IOC. All planned/targeted as of mid-2026. DefenseScoop, June 3, 2026.
The Trace Systems MPE-S contract carried a $998 million single-award IDIQ ceiling, executed by the Air Force Mission Partner Capabilities Office, period of performance September 2019 to September 2024, supporting about 45,000 users. A ceiling is a maximum orderable, not obligated or spent dollars; this vehicle is past, not current. Trace Systems, MPE-S.
GAO-25-106454 (2025): DoD requested more than $1.4 billion for CJADC2 activities in its FY2025 budget request (a request, not spending, and not MPE-specific); MPE is listed among EUCOM data-sharing systems; capabilities are "developed... largely in isolation" without a comprehensive framework. GAO-25-106454.
GAO-16-593 (2016): DoD needed to strengthen governance and management of the Joint Information Environment against leading practices for scope, cost, schedule, workforce, and security. JIE is distinct from MPE; used as precedent for the recurring governance pattern. GAO-16-593.
CENTRIXS-ISAF and CENTRIXS-GCTF were integrated into the CENTCOM Partner Network (CPN); medium confidence on the specific mapping, sourced to a specialist blog citing an army.mil article that could not be re-read directly. Electrospaces.net.
CPN-X resolves to "CENTCOM Partner Network Extended," a SECRET RELEASABLE coalition network listed in a USAF FY2021 budget justification; read via a third-party mirror, with no dollar figure broken out. It is not the Army WIN-T "Command Post Node," and "Coalition Partner Network" is a mistaken secondary expansion. USAF FY2021 budget justification (Stratvocate mirror).
Design rationale (separate, non-interconnecting networks by releasability) corroborated in a Naval Postgraduate School analysis. NPS Calhoun.

Coalition networks: CENTRIXS, the Mission Partner Environment, and the sprawl the Pentagon keeps trying to untangle