GCRI · Foundations
Edition Core: the temporary technical backbone assembled each year for Nexus Universe
Each year the programme assembles a temporary technical environment from contributed computing, networks and teams, tests it for a week and stands it down. This page explains how it is built, governed and measured, and what it is not.
- As of
- 21 October 2026
- 読む
- 12 minutes
- エビデンス
- 18 sources, cited in order
手法とエビデンス
- エビデンス
- 18 sources, cited in order and listed under References
- 数字
- 6 figures
- Data availability
- Every chart's data is published as a table on this page.
01What the edition Core is
What the edition Core is
The edition Core is the temporary technical environment that the Nexus Consortium assembles each year from contributed, authorised computing, networks and teams for the week of controlled exercises of Nexus Universe, and then stands down 1. The Nexus Consortium is the public-good programme through which The Global Centre for Risk and Innovation (GCRI), The Global Risks Forum (GRF) and The Global Risks Alliance (GRA) help countries examine interconnected risk with shared methods. GCRI, which publishes the design on therisk.global, coordinates both layers of shared computing, the Core and the durable Nexus Network, through GCRI Canada 1.
The Core is not a system of its own and not a product. It is an assembly of capacity governed by the systems that are the authority for each part of it: NXS-COR (compute, nodes and runtimes) for placement, capacity and attestation, NXS-NET (network topology and continuity) for links and failure domains, NXS-FAC (blueprints, builds and signing) for signed builds and NXS-NSF (trust, authorization and assurance) for policy and keys 8,9,10,11. An earlier doctrine used the name for a national secure-processing layer; in the current model that work belongs to those systems, and the edition Core names only the annual backbone.
GCRI states plainly what the two layers are not: the Network is not a permanent global supercomputer, and the Core is not installed, standing capacity 1.
02Why an annual programme needs a temporary stack
Why an annual programme needs a temporary stack
An exercise that tests how a grid failure reaches a hospital, a payment system and a port needs the models, data, networks and computing of several institutions working together for a short time, under control. Building that as permanent shared infrastructure would be slow, expensive and a concentration of risk in its own right; improvising it would produce results nobody could reproduce. A temporary stack, assembled to a plan, measured, used and then closed, sits between those failures.
The cost of computing makes the discipline sharper. In the United States, data centres used about 4.4% of national electricity in 2023, up from 58 TWh in 2014 to 176 TWh in 2023, and the Department of Energy’s December 2024 report projects 325 to 580 TWh by 2028, or about 6.7% to 12% of national use 7. Capacity is an energy, water and grid question before it is a technical one, which is why the programme counts the footprint of anything it assembles, keeping measurements apart from estimates 1.
The temporary form also protects the programme’s neutrality. Because nothing persists by default, a contributor gains no standing position: an offer confers no governance role, and allocation is a separate, approved step 1.
Fig. 04Computing capacity is also an energy question
Figure 4 in words, and its data
Left panel, terawatt-hours a year on a scale to 600: 2014, 58; 2023, 176, highlighted; 2028, a projected range from 325 to 580 shown as a solid lower bar and a hatched upper range. Right panel, per cent of total US electricity on a scale to 12: 2023, about 4.4, highlighted; 2028, a projected range from 6.7 to 12. The data are tabulated below the figure.
| 年 | Electricity use | Share of US electricity |
|---|---|---|
| 2014 | 58 TWh 7 | Not stated in the release |
| 2023 | 176 TWh 7 | About 4.4% 7 |
| 2028 (projected) | 325 to 580 TWh 7 | About 6.7% to 12% 7 |
03Persistent, temporary and durable
Persistent, temporary and durable
Shared computing in the programme has three layers, and each keeps something different 1. The persistent coordination layer holds what continues from year to year: authorised institutions and projects, resource catalogues and qualification, identity, rights and methods, and retained evidence and publication records. The temporary technical environment is the Core itself: qualified networking and high-performance computing, accelerators, cloud and edge capacity, secure data spaces and cyber ranges, simulation environments and technical support. The durable outputs are what remain once the Core is closed: trained teams and tested configurations, reproducible workloads and software, datasets where permitted and evaluation results, and improvements to institutional practice 1.
The Nexus Network is the durable federation of retained national and regional sites 1. Its annual review is part of the fourth quarter of the programme year, which publishes measured capacity results and resource qualification changes and draws on measured results only; device counts are not results 12.
Fig. 01Persistent, temporary and durable
Figure 1 in words
Band one, persistent coordination: authorised institutions and projects, resource catalogues and qualification, identity, rights and methods, retained evidence and publication records. Band two, highlighted, the temporary technical environment, the edition Core: qualified networking and high-performance computing, accelerators, cloud and edge capacity, secure data spaces and cyber ranges, simulation environments and technical support. Band three, durable outputs: trained teams and tested configurations, reproducible workloads and software, datasets where permitted and evaluation results, improvements to institutional practice. A side panel: the Nexus Network is not a permanent global supercomputer; the Core is not installed, standing capacity; neither membership nor sponsorship secures an allocation.
04From offer to accepted capacity
From offer to accepted capacity
Computing centres, research networks, telecoms, cloud and equipment providers can offer qualified resources, and technology teams can answer testbed and mission calls; selection follows documented criteria, evaluation never depends on contribution size, and testing is not certification 4. An offer records its owner and operating contact, its legal and use constraints, location and availability window, its hardware, software, interconnect, memory and storage profile, its security state, supported workloads and measured limits, and its return and sanitisation requirements 1.
A contributed resource then moves through seven states: offered, qualified, reserved, allocated, accessible, measured and accepted. None implies the next. An offer is never counted as reserved, consumed or funded, and neither membership nor sponsorship secures an allocation 1. The offering that frames this work, Annual Universe Missions and Contributed Infrastructure, sets the same boundary and lists its deliverables: a mission contract, a resource allocation, Core Build evidence, exercise evidence, a technical closeout and an improvement backlog 13. Providers start at technology and infrastructure providers.
Fig. 02Seven states, none of which implies the next
Figure 2 in words
Seven bars of increasing height. One, offered: receives a disposition and confers no governance role. Two, qualified: a qualification record if suitable. Three, reserved: never assumed from an offer. Four, allocated: a separate, approved step. Five, accessible: still not counted as capacity. Six, measured: with its stated conditions. Seven, accepted, highlighted: counted only once measured and accepted. Below, an offer records owner and operating contact; legal and use constraints, location and availability window; hardware, software, interconnect, memory and storage profile; security state, supported workloads and measured limits; and return and sanitisation requirements.
05Built in a month, tested in a week, stood down in a day
Built in a month, tested in a week, stood down in a day
The Core follows the programme’s time scales 5. A year of preparation ends with an approved mission portfolio, qualified inputs and an achievable build plan, through gates G0 to G2. In one month the technical office assembles the Core, binds authorised resources, integrates software and data, tests security and interoperability and rehearses; the exit is a Core Build Report with measured readiness, known defects and go or no-go decisions, and gate G3 accepts the build for exercise 5. In one week of controlled exercises, scenario work and benchmarks, gate G4 authorises controlled live operation 5,12.
The proposed one-day stand-down closes new submissions, checkpoints authorised jobs, freezes the evidence manifest, revokes temporary credentials, reconciles equipment and records open exceptions; it becomes a commitment only when the annual programme records it, and review, reconciliation and corrections continue afterwards 1,5. Gate G5 records the evidence seal and the stand-down. No promise is made that data is deleted within a day: media are purged or retained under approved, validated rules 1, the programme-level posture of NIST SP 800-88 Revision 2 of September 2025 14.
Fig. 03Prepared for a year, built in a month, tested in a week
Figure 3 in words
One year, prepare, gates G0 to G2: national demand, research, participation, monthly working-group delivery, quarterly evidence and resource qualification; exit an approved mission portfolio and a build plan. One month, build, gate G3, highlighted: assemble the Core, bind authorised resources, integrate software and data, test security and interoperability, rehearse; exit Core Build Report, measured readiness, known defects, go or no-go. One week, act and test, gate G4: controlled exercises, scenarios, benchmarks, independent observation and leadership deliberation; exit run, decision and observation records and an evidence seal. One day, stand down, gate G5, dashed as proposed: stop new work, checkpoint jobs, revoke temporary access, release capacity, seal evidence. Following period, review and renew, gates G6 and G7: validate archives, complete review, publish approved results, close corrective actions.
06Who builds, who governs, who decides
Who builds, who governs, who decides
GCRI’s technical office for Nexus Universe does six things: mission scoping, qualification of data, methods and contributed resources, the Core build with its build evidence and known defects, support to exercise controllers and independent evaluators, evidence capture with run, decision and observation records sealed for review, and a technical closeout with an improvement backlog 2. It works through the year, not only in the week of exercises, and its outputs are records that governance, evaluators and reviewers can inspect 2.
GRF holds four decisions within its instruments: the annual mandate, mission admission, the gate decisions G0 to G7 and what is published, with the director who admits participants kept apart from controllers and evaluators 3. In GRF’s words, the technical office of GCRI "builds and operates the temporary edition Core for each edition", qualifies data and methods and captures evidence, while GRF governs the programme; the technical office qualifies the technical build and does not admit missions 3. Finance and insurance readiness is a planned track of Action Week, which GRA is designed to run through its Universe finance-readiness programme, with diligence questions and protection gaps, never an approval or a transaction 15,3.
Hosting is not ownership: a host gains no control of the programme, its records or its results, and hosts and venues have their readiness evidenced before any event is presented as confirmed 3,2. Decisions and execution outside the programme remain with the relevant government, operator or licensed institution 3.
07Each system answers for its own records
Each system answers for its own records
The Core is governed by dividing authority, not by concentrating it. NXS-COR is the single owner of runtime placement capacity and attestation; it keeps workload requirements, node observations with their trust state, placement constraints and maintenance windows, and its method boundary warns that a feasible simulation is not proof a node is reachable or attested 8. NXS-NET keeps link observations with their collection method and failure domains, and does not treat a simulated alternate route as a tested communications failover 9. NXS-FAC approves the exact artefact digest rather than a mutable branch, keeps signing credentials outside the room, and does not treat a signed build as permission to deploy 10.
NXS-NSF holds policy, consent, delegation, the trust root and key custody; key material stays in KMS or HSM in the sovereignty cell, and a revocation invalidates dependent decisions and cached grants 11. NXS-GRX (canonical data, provenance and risk) holds the evidence that the Core produces, with source releases, quality envelopes and correction lineage 16. Each system exports read projections through typed contracts, and no other system writes its records 8.
Fig. 05Each system answers for its own records
Figure 5 in words
NXS-COR, compute, nodes and runtimes: authority for runtime placement, capacity and attestation, workload requirements and maintenance windows; boundary, a feasible simulation is not proof a node is reachable or attested. NXS-NET: network and service boundaries, link observations, failure domains and change windows; a simulated alternate route is not a tested failover. NXS-FAC: build signing, source snapshot, build contract, verification suite and artefact digest; a signed build is not permission to deploy. NXS-NSF: policy, consent, delegation, trust root and key custody; key material stays in KMS or HSM in the sovereignty cell. NXS-GRX: source releases, fusion recipes, quality envelopes and correction lineage; observation, inference, prediction and simulation are never flattened into one truth value.
08How capacity is measured and described
How capacity is measured and described
Claims about computing are among the easiest to inflate, and the programme publishes four rules for describing capacity 1. State the conditions: measured capability is published with its workload, data size, hardware, topology, precision, concurrency, software version and interval. Separate the numbers: peak, sustained, available and allocated capacity are kept apart, and hardware specifications are never summed into a performance claim. Federation is not one machine: local low-latency computing and wide-area federation are described separately. Count the footprint: energy, cooling, water, transfers and logistics are tracked, with measurements kept apart from estimates 1.
Capacity counts only once it is measured and accepted, and the Core Build Report records measured readiness alongside known defects rather than a single score 1,5. The programme states no readiness percentage for any part of the Core: no cycle has run 6.
Fig. 06Rules for describing capacity
Figure 6 in words
Rule one, state the conditions: publish measured capability with workload, data size, hardware, topology, precision, concurrency, software version and interval. Rule two, highlighted, separate the numbers: keep peak, sustained, available and allocated capacity apart and never sum hardware specifications into a performance claim. Rule three, federation is not one machine: describe local low-latency computing and wide-area federation separately. Rule four, count the footprint: track energy, cooling, water, transfers and logistics, keeping measurements apart from estimates.
09Instruments the stack is designed against
Instruments the stack is designed against
The design cites established technical guidance as references, not as claims of conformance. Access is designed on NIST SP 800-207, Zero Trust Architecture, so that no trust is granted by network location alone 17. Builds follow the practices of the Secure Software Development Framework, NIST SP 800-218, with pinned inputs, declared nondeterminism and independent rebuild comparisons in NXS-FAC 18,10. Media sanitisation at stand-down follows NIST SP 800-88 Revision 2 14.
| 年 | 公開 | Where it applies |
|---|---|---|
| 2020 | NIST SP 800-207, Zero Trust Architecture | Per-request access decisions under NXS-NSF 17,11 |
| 2022 | NIST SP 800-218, SSDF Version 1.1 | Build practices and signing in NXS-FAC 18,10 |
| 2025 | NIST SP 800-88 Rev. 2, Guidelines for Media Sanitization | Return and sanitisation of contributed equipment at stand-down 14,1 |
| 提案中 | Nexus Risk Standards system profiles NRS-S4 (NXS-COR) and NRS-S10 (NXS-NET) | Status OPEN 8,9 |
10What is changing
What is changing
- Compute meets the grid. With US data-centre demand projected at 6.7% to 12% of national electricity by 2028, any temporary assembly of computing has to account for grid capacity and footprint 7.
- Signed, reproducible builds. Approving an exact artefact digest, not a branch, is becoming the basic test that a build used in an exercise is the build that was reviewed 10,18.
- Placement as constrained optimisation. Placement over verified capacity and policy, with oversubscription modelled explicitly, replaces placement by availability alone 8.
- Sanitisation as a programme. NIST’s 2025 revision treats sanitisation as a programme matched to data sensitivity, which suits equipment that returns to its owners after stand-down 14.
11Three faces of the Core
Three faces of the Core
GCRI builds and operates the Core and answers for its evidence, through the technical office and GCRI Canada 2,1. GRF decides whether an edition runs, which missions enter and what is published, and keeps hosting and sponsorship from conferring control, as set out in programme governance 3. GRA reads what the exercises mean for finance and insurance without any transaction taking place 15. The annual operating cycle at GRF and the annual programme cycle at GCRI describe the same year from the two sides.
What this page does not claim: no Core has been built, no allocation is open and no edition is on record; the programme is at service level R0. An offer of capacity is not an allocation, a qualification record is not certification, and a signed build is not permission to deploy. GCRI does not regulate or certify, GRF does not regulate, finance or procure, and GRA does not lend, underwrite, rate or invest. Technology and infrastructure providers work with the programme at arm’s length under written agreements and hold no role in its stewardship.
質問
質問と回答
Is any edition Core capacity available now?
No. No allocation is open, and the current service level is information and enquiries. A provider can make an enquiry about a resource offer, which receives a disposition and, if suitable, a qualification record; allocation is a separate, approved step.
Does contributing computing give my organisation a role in the programme?
No. An offer confers no governance role, neither membership nor sponsorship secures an allocation or mission admission, and evaluation never depends on the size of a contribution. Hosting is not ownership of the programme, its records or its results.
How is the edition Core different from NXS-COR?
The edition Core is the temporary backbone assembled for one edition and then stood down. NXS-COR is one of the twenty Nexus systems, the authority for compute, nodes, runtimes and placement, and it governs the computing part of the Core along with NXS-NET, NXS-FAC and NXS-NSF.
Is data deleted when the Core is stood down?
No such promise is made. Temporary credentials are revoked and the evidence manifest is frozen at the proposed stand-down; media are then purged or retained under approved, validated rules, and review, reconciliation and corrections continue afterwards.
Is the Core a supercomputer?
No. The Core is not installed, standing capacity, and the Nexus Network is not a permanent global supercomputer. Local low-latency computing and wide-area federation are described separately, and hardware specifications are never summed into a performance claim.
利用条件
用語集
- edition Core
- The temporary technical backbone assembled each year from contributed, authorised computing, networks and teams for Nexus Universe, then stood down.
- Nexus Network
- The durable federation of retained national and regional sites of shared computing.
- Resource states
- Offered, qualified, reserved, allocated, accessible, measured and accepted; none implies the next.
- コア構築レポート
- The exit record of the one-month build: measured readiness, known defects and go or no-go decisions, before gate G3.
- 警戒解除
- The proposed one-day closing of the Core: no new work, jobs checkpointed, temporary access revoked, evidence sealed.
関連
Further reading across the network
Reuse and cite
Text CC BY 4.0. Cite the page with its date; data tables are linked under each figure.
出典
参考文献
18 sources, numbered in the order they are first cited
- The Global Centre for Risk and Innovation (2026-10). Network and Core. therisk.global. Accessed 2026-10-21.
- The Global Centre for Risk and Innovation (2026-10). Nexus Universe technical office. therisk.global. Accessed 2026-10-21.
- The Global Risks Forum (2026-10). Programme governance (Nexus Universe). globalriskforum.com. Accessed 2026-10-21.
- The Global Centre for Risk and Innovation (2026-10). Technology and infrastructure providers. therisk.global. Accessed 2026-10-21.
- The Global Centre for Risk and Innovation (2026-10). Annual programme cycle. therisk.global. Accessed 2026-10-21.
- The Global Centre for Risk and Innovation (2026-10). Claims and readiness register. therisk.global. Accessed 2026-10-21.
- U.S. Department of Energy (2024-12-20). DOE Releases New Report Evaluating Increase in Electricity Demand from Data Centers. www.energy.gov. Accessed 2026-10-21.
- Nexus documentation (docs.therisk.global) (2026-10). NXS-COR: compute, nodes, runtimes, placement and infrastructure lifecycle. docs.therisk.global. Accessed 2026-10-21.
- Nexus documentation (docs.therisk.global) (2026-10). NXS-NET: network topology, connectivity and communications continuity. docs.therisk.global. Accessed 2026-10-21.
- Nexus documentation (docs.therisk.global) (2026-10). NXS-FAC: blueprints, builds, verification, signing and assembly. docs.therisk.global. Accessed 2026-10-21.
- Nexus documentation (docs.therisk.global) (2026-10). NXS-NSF: trust, authorization, mandate, policy and assurance. docs.therisk.global. Accessed 2026-10-21.
- Nexus documentation (docs.therisk.global) (2026-10). The programme year. docs.therisk.global. Accessed 2026-10-21.
- Nexus documentation (docs.therisk.global) (2026-10). Status and service levels (offering O09, Annual Universe Missions and Contributed Infrastructure). docs.therisk.global. Accessed 2026-10-21.
- U.S. National Institute of Standards and Technology (NIST) (2025-09-26). SP 800-88 Rev. 2, Guidelines for Media Sanitization. csrc.nist.gov. Accessed 2026-10-21.
- The Global Risks Alliance (2026-10). Universe finance-readiness programme. globalriskalliance.com. Accessed 2026-10-21.
- Nexus documentation (docs.therisk.global) (2026-10). NXS-GRX: canonical data, provenance, risk and federation. docs.therisk.global. Accessed 2026-10-21.
- U.S. National Institute of Standards and Technology (NIST) (2020-08). SP 800-207, Zero Trust Architecture. csrc.nist.gov. Accessed 2026-10-21.
- U.S. National Institute of Standards and Technology (NIST) (2022-02). SP 800-218, Secure Software Development Framework (SSDF) Version 1.1. csrc.nist.gov. Accessed 2026-10-21.
記録
Editorial record
Facts and programme states are given as of 21 October 2026. This page is reviewed at least every 180 days and whenever a programme state, a service level or a cited figure changes. It replaces an earlier version of the same post, which presented the Core as a stack of named modules operated with platforms under names that are now retired, and described sponsor-supported infrastructure and cyber ranges without the resource states, gates and measurement rules now on record.
Published by the GCRI editorial desk under the scientific and editorial integrity policy. To report an error or request a correction, use the corrections and concerns route; every accepted correction is recorded as a new version, never a silent edit.
Cite this article
GCRI editorial desk (2026, June 8). Edition Core: the temporary backbone of Nexus Universe. The Global Centre for Risk and Innovation. https://therisk.global/from-temporary-technical-infrastructure-to-nexus-core-designing-a-mission-grade-stack-for-nexus-universe/
@online{gcri2026fromtemporarytechnical,
author = {{GCRI editorial desk}},
title = {{Edition Core: the temporary backbone of Nexus Universe}},
publisher = {The Global Centre for Risk and Innovation},
year = {2026},
month = {6},
url = {https://therisk.global/from-temporary-technical-infrastructure-to-nexus-core-designing-a-mission-grade-stack-for-nexus-universe/},
language = {en}
} TY - ELEC AU - GCRI editorial desk TI - Edition Core: the temporary backbone of Nexus Universe PB - The Global Centre for Risk and Innovation PY - 2026 DA - 2026/06/08 UR - https://therisk.global/from-temporary-technical-infrastructure-to-nexus-core-designing-a-mission-grade-stack-for-nexus-universe/ LA - en ER -
Republish
Republish or translate this article free of charge, in print or online, under CC BY 4.0: keep the byline and The Global Centre for Risk and Innovation, link to the original and note any change. Figures, maps and twins made by the institutions carry the same licence; third-party data keeps its own credit line.
Attribution to paste
<p>Edition Core: the temporary technical backbone assembled each year for Nexus Universe — GCRI editorial desk, The Global Centre for Risk and Innovation. First published 8 June 2026 at <a href="https://therisk.global/from-temporary-technical-infrastructure-to-nexus-core-designing-a-mission-grade-stack-for-nexus-universe/">https://therisk.global/from-temporary-technical-infrastructure-to-nexus-core-designing-a-mission-grade-stack-for-nexus-universe/</a>. Republished under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/).</p>