Research libraryFactories on a Weak Grid
Chapter 02 of 14 · Part I · Draft web edition
A useful design envelope records the duration, coincidence and recovery of power and network failures at the intended site.
The cloth is woven from this chapter's 4 exhibits: each weft is one exhibit's own data, normalised; each warp group is a section. The probe reads the exhibit's own value where it crosses. Procedural visualisation, not a measurement.
Read-out
World Bank Cameroon 2024 Country Profile, printed page 6 Figure 8 and page 11 Appendix. Survey population covers 615 formal non-agricultural private firms (5+ employees). Frequency does not measure outage duration or feeder tail risk.
Unit: outages in a typical month. Status: sourced. Sources: S01.
Overture · 1 min
National enterprise surveys establish that electrical instability is a material operating constraint across Cameroon, where 93.3 percent of surveyed formal firms experienced power outages and reported 10.4 outages in a typical month.[S01] Yet an average monthly outage count cannot size an uninterruptible power supply (UPS), dimension an offline event buffer or verify whether a candidate backhaul link survives when utility mains drop.[S01] [S05] GFT argues that before an industrial sponsor selects edge hardware, battery autonomy or synchronisation intervals for a textile plant, the project team must replace macroeconomic averages with a time-synchronised site survey that logs the duration, coincidence and recovery lag of electrical and network failures at the actual plant boundary.[S03] [S05] Until those feeder-level and carrier-level observations are collected under a documented protocol, every autonomy duration, queue budget and availability figure remains a conditional design scenario rather than an observed site fact.[S02] [S05]
2.1
Related · chapter 016 paragraphs step it
Read-out
World Bank Cameroon 2024 Country Profile, published 2025, printed page 6, Figure 8. Comparator groups and vintages are as published, not a textile-only or site-level forecast.
Unit: percent of firms experiencing electrical outages. Status: sourced. Sources: S01.
Read this exhibit in its chapter ↗Macroeconomic business-environment indicators justify why digital resilience belongs in the capital-expenditure scope of a Cameroonian factory, while simultaneously demonstrating why aggregate survey numbers cannot serve as engineering inputs.[S01] [S02] In the World Bank Enterprise Survey for Cameroon 2024, field teams interviewed business owners and top managers across 615 formal private non-agricultural firms employing at least five workers between June 2024 and February 2025.[S01] Across that national sample, 93.3 percent of firms reported experiencing electrical outages, compared with 69.7 percent for the Sub-Saharan Africa comparator group and 55.0 percent for the lower-middle-income comparator group.[S01] When respondents selected the single biggest obstacle to their business from 15 constraints, 12.1 percent of all firms and 25.3 percent of large firms (100 or more employees) identified electricity as their primary obstacle, ahead of political instability at 19.7 percent and access to finance at 11.3 percent.[S01]
The survey also reports a second, distinct reliability indicator: the number of electrical outages experienced in a typical month, presented in Exhibit 5.[S01] Surveyed firms in Cameroon reported 10.4 electrical outages in a typical month, against 6.9 in the Sub-Saharan Africa comparator and 5.3 in the lower-middle-income comparator.[S01] By establishment size, outage incidence reached 95.7 percent among small firms (5 to 19 employees), 90.6 percent among medium firms (20 to 99 employees) and 77.9 percent among large firms (100 or more employees), whereas typical monthly frequency stood at 10.6 outages for small firms, 9.4 for medium firms and 11.8 for large firms.[S01] Firms applying for an electrical connection waited an average of 35.5 days across all surveyed establishments.[S01]
As Exhibit 5 shows, the published frequency of 10.4 outages per month is nearly double the lower-middle-income benchmark of 5.3 outages per month.[S01] However, an engineering team must separate what Exhibit 5 proves from what it cannot establish.[S01] First, a monthly count of 10.4 events carries no information about the distribution of event durations.[S01] Ten interruptions lasting three minutes each impose 30 minutes of cumulative monthly downtime in an illustrative calculation and require only short ride-through storage, whereas ten interruptions containing two eight-hour feeder faults impose a completely different battery and data-retention duty.[S05] Second, the Country Profile headline chart reports incidence and monthly count on separate axes and does not publish a site-level duration percentile, phase-imbalance count or voltage-sag profile.[S01]
The World Bank Enterprise Surveys methodology explains why national point estimates cannot be assigned to a specific industrial plot or textile mill.[S02] Under the global methodology applied since 2005 to 2006 and updated in August 2022 to use International Standard Industrial Classification (ISIC) Revision 4 codes, the formal-sector survey samples registered companies with five or more workers and at least 1 percent private ownership across manufacturing (ISIC codes 10 to 33), construction, retail, wholesale, hospitality, transport, storage, communications, information technology, and professional, scientific and technical activities.[S02] Agriculture, fishing, mining, public utilities, financial intermediation, public administration, education, health and 100 percent state-owned enterprises are excluded.[S01] [S02]
In the Cameroon 2024 sample of 615 establishments, only 161 firms belonged to manufacturing (49 food-manufacturing and 112 other-manufacturing firms), alongside 126 retail and 328 other-services firms, stratified across the Centre (207 firms), Littoral (199), North (146) and West (63) regions.[S01] Furthermore, because the World Bank tracks panel enterprises across survey waves, sampling weights are adjusted to account for altered probabilities of inclusion in the sample frame.[S02] Even within manufacturing, outage incidence reached 100.0 percent in food manufacturing and 87.3 percent in other manufacturing, with typical monthly counts of 11.9 and 10.4 outages respectively.[S01]
Consequently, the 10.4 monthly outage figure is a sampling-weighted, multi-sector national mean, not a textile-sector measurement or a feeder log for Douala, Dibamba or Garoua.[S01] [S02] While Cameroon's cotton-to-garment opportunity rests on moving local fibre processing from roughly 5 percent (with about 95 percent exported raw) toward the national 50 percent target by 2030, and while historic reports on CICAM cited July 2023 capacity utilisation of 33.7 percent at Garoua, 48.9 percent at CICAM 1 and 10.83 percent at CICAM 2, neither source provides feeder-specific electrical or telecoms telemetry.[R06] [R07] [R12] Treating the national 10.4 monthly average as a factory design envelope would conflate a macroeconomic statistic with a site engineering specification.[S01] [S02]
2.3
A conventional utility billing meter or monthly kilowatt-hour totaliser measures cumulative energy transfer over a billing cycle, which is useful for tariff reconciliation and resource accounting but blind to the transient disturbances that disrupt digital systems.[S03] [S05] Operational technology (OT) and edge-computing equipment, including programmable logic controllers, machine-vision cameras, industrial switches, time servers and local database hosts, fail on timescales of milliseconds to seconds when exposed to voltage sags, single-phasing, frequency excursions or slow contactor transfer intervals.[S03] A plant energy meter can record normal daily consumption while a 400-millisecond illustrative voltage dip or an uncoordinated generator changeover reboots an unprotected switch and severs a transaction in flight.[S03] [S13]
…normal daily consumption while a 400-millisecond illustrative voltage dip or an uncoordinated generator changeover…
S03, S13 · illustrative
To specify the continuity boundary required by NIST Special Publication 800-82 Revision 3 for operational technology reliability and safety and by NIST Special Publication 800-34 Revision 1 for information-system contingency planning, the site baseline must capture a discrete power-event log rather than a simple energy total.[S03] [S05]
Distinguishing brownouts, single-phasing and transfer intervals from complete blackouts is essential for textile operations.[S03] During changeover from utility mains to a site diesel generator through an open-transition automatic transfer switch, the motive-power bus experiences a dead interval of several seconds to several minutes while the generator starts and stabilises voltage and frequency.[S03] [S05] If an edge gateway or network switch sits upstream of an online UPS, that open-transition transfer interval reboots the digital node just like an eight-hour illustrative grid blackout.[S05] Moreover, although the Linux ext4 filesystem enables write barriers by default (barrier=1) to enforce on-disk journal ordering and defaults to a five-second commit interval (commit=5), a sudden loss of DC power on an unbacked host can still lose up to five seconds of un-fsynced metadata, and SQLite databases in write-ahead logging (WAL) mode with PRAGMA synchronous=NORMAL can roll back recently committed transactions after a power loss.[S16] [S29]
A site power-event log defines the disturbance environment that the digital system must survive, but it does not replace qualified electrical engineering design.[S03] GFT's recording method specifies what telemetry IT and OT architects need to classify services and size digital backup autonomy; it does not prescribe earthing topologies, short-circuit ratings, arc-flash boundaries or protective relay settings.[S03] The project sponsor must commission a licensed electrical engineer to inspect the site transformer, verify neutral-to-earth bonding, specify overcurrent and surge protection, and certify that any temporary power-quality logger is installed safely.[S03]
2.5
Just as a monthly energy meter hides power transients, a router link light or a simple Internet Control Message Protocol (ICMP) echo test hides the network failure modes that stall industrial data synchronisation.[S18] [S24] In practice, an edge outbox attempting to forward committed production events to a central repository depends on a multi-layer protocol chain, and failure at any intermediate layer renders the backhaul unusable for application recovery.[S12] [S23] [S24]
SYN, SYN-ACK, ACK).[S24] A bare TCP connection proves only that three small control segments traversed the path; it does not prove bulk payload capacity or remote application health.[S24]200 OK or 201 Created carrying the committed event ID) across the open TLS session before disconnection.[S12] [S13] [S23]To measure usable connectivity without inventing carrier performance figures, the site survey should deploy a lightweight software probe on the candidate site network that executes this six-step chain against a test endpoint at a regular cadence during agreed production hours, testing representative payload sizes under RFC 6349 principles.[S05] [S0202] [S0203] Because no site carrier trace has been collected for a GFT textile plant in Cameroon as of the edition date, this paper publishes zero Cameroon bandwidth, latency or packet-loss claims.[S18]
Three rules govern the connectivity probe log.[S10] [S22] [S0203] First, the probe must record its own state using persistent local logging such as systemd-journald with Storage=persistent so records survive reboots.[S22] Second, the probe host must log clock synchronisation status and offset under RFC 5905 and RFC 8633 (chronyd with rtcsync or rtcfile).[S10] [S11] [S30] Third, whenever the probe host is unpowered or isolated from the LAN, the interval must be flagged explicitly as a monitoring gap (observer_offline), never interpolated as available or failed.[S05] [S0203]
Interlude · the cloth half-woven
Woven: sections 2.0 to 2.5Still to weave: 2.7 to 2.12
2.7
Exhibit 066 paragraphs step it
Read-out
Illustrative synthetic 160-hour (four-week, single-shift) site log demonstrating the joint-state matrix method. Values are assumed scenario counts that sum to 160 hours per row; they are not observed Cameroon statistics.
Unit: illustrative shift-hours out of 160. Status: illustrative. Sources: S05, S18. Illustrative figures are teaching examples, not measured factory results.
Even when a project team measures site electricity and site connectivity with proper instruments, analysing the two logs in isolation produces a dangerous sizing error.[S03] [S05] Reliability engineers who multiply a standalone mains-availability fraction by a standalone WAN-availability fraction implicitly assume that electrical outages and telecommunications outages are statistically independent.[S05] In industrial zones, that assumption fails because power feeders and telecoms networks share physical and electrical dependencies:[S03] [S18]
To make these interactions visible, the site baseline must cross-tabulate power state and network state over identical time slices to form a joint outage envelope, illustrated in Exhibit 6.[S05] [S18]
In Exhibit 6, three illustrative site topologies are compared across a synthetic 160-hour production window in which utility mains are unavailable for 20 hours (12.5 percent of shift hours, an illustrative scenario input).[S05] [S18] In the first row (Unbacked CPE router), because the site modem loses power whenever utility mains fail, the system spends zero hours in Mains Down / WAN OK and 20 hours in the joint failure state (Mains Down / WAN Down), bringing total WAN unavailability to 24 hours out of 160.[S03] [S05] In the second row (Shared-feeder tower battery), protecting the plant router on a local UPS yields 8 hours of connected operation during short cuts, but the external tower shares the utility feeder and exhausts its battery during longer cuts, leaving 12 hours in Mains Down / WAN Down.[S05] [S18] Only in the third row (Independent dual-path target), where the router is protected and secondary backhaul avoids the local feeder failure mode, does the plant remain connected for 18 of the 20 mains-outage hours.[S05] Exhibit 6 is methodological: it shows why separate marginal averages hide the exact cell (Mains Down / WAN Down) that governs simultaneous battery and storage duty.[S02] [S05]
Each of the four cells in Exhibit 6 places a distinct demand on the continuity architecture.[S03] [S05] Mains OK / WAN OK represents normal connected operation.[S08] Mains OK / WAN Down represents isolated operation while machinery runs at full speed on utility power, generating production events at peak shift rate and governing maximum local queue intake.[S03] [S08] Mains Down / WAN OK allows a battery-backed digital node to flush final transactions and send a power-loss alert before shutting down or switching to generator power.[S05] [S14] Mains Down / WAN Down represents compound isolation, where all point-of-work capture must commit atomically to local storage without external DNS, NTP, identity or cloud dependencies.[S08] [S10] [S13]
2.8
Averages conceal the extremes that break bounded queues and exhaust batteries.[S05] [S08] Whether a local SQLite WAL partition fills up or a protected digital circuit brown-outs depends on the upper tail of the outage-duration distribution: the longest single offline interval and the longest cluster of back-to-back outages whose recovery gap is too short to recharge batteries or drain queues.[S05] [S08] [S16]
To characterise outage durations without inventing parametric curves, the site survey should compile observed durations into an empirical distribution function (EDF) as defined in RFC 2330.[S0203] For $N$ scalar duration measurements, $F(x)$ gives the fractional proportion of measurements less than or equal to $x$, and the $p$-th percentile is the smallest observed value $x$ for which $F(x) \ge p$ percent.[S0203]
RFC 2330 states a strict mathematical rule for empirical percentiles: any percentile summary must report the sample size $N$, because any percentile difference finer than $1/N$ is below the resolution of the sample.[S0203] If a four-week site survey observes $N = 10$ electrical outages (near the World Bank Cameroon 2024 typical monthly count of 10.4 outages), the sample resolution is $1/10 = 10$ percentage points.[S01] [S0203] On 10 events, claiming to distinguish a 95th percentile from a 99th percentile is an interpolation artefact, and even a 90-day baseline with an illustrative $N = 30$ outages ($1/30 \approx 3.33$ percent resolution) cannot resolve a true 99th percentile.[S0203]
> 4.0 h, censored), never dropped or recorded as a completed 4.0-hour event.[S05]Because pre-commissioning samples are small, the design offline interval (such as the illustrative two-hour protected digital autonomy in Chapter 4 and eight-hour event retention in Chapter 7) must be set as an explicit management risk-acceptance boundary informed by observed maxima, censored events and operator interviews, never extrapolated from the World Bank's 10.4 monthly frequency figure.[S01] [S05]
2.10
Mains voltage returns
0: Mains voltage returns. Utility feeder re-energises at t = 0 min; ATS waits for voltage stability. · illustrative
An external utility or telecommunications outage is shorter than the operational disruption it causes.
Stable power and CPE boot
2: Stable power and CPE boot. Transfer switch closes; unbacked ISP modem and access switch cold-boot. · illustrative
Link up and DNS recovery
5: Link up and DNS recovery. Carrier link establishes; negative DNS cache entries expire and hostname resolves. · illustrative
TCP and TLS 1.3 handshake
7: TCP and TLS 1.3 handshake. Edge client completes TCP 3WHS, TLS 1.3 certificate validation and auth. · illustrative
First remote durable commit
8: First remote durable commit. Cloud endpoint accepts first replayed batch and returns committed-ID receipt. · illustrative
Backlog clearance complete
45: Backlog clearance complete. Outbox drains accumulated offline events while live shift arrivals continue. · illustrative
Illustrative event sequence in relative minutes from physical voltage return (t = 0). Timings are hypothetical scenario values showing protocol and operational ordering under RFC 9293 (TCP), RFC 9846 (TLS 1.3) and NIST SP 800-184; they are not measured site timings.
Unit: illustrative relative minutes from mains return. Status: illustrative. Sources: S06, S23, S24. Illustrative figures are teaching examples, not measured factory results.
2.10 · the bodyExhibit 07 above; its figures are illustrative where marked.
An external utility or telecommunications outage is shorter than the operational disruption it causes.[S06] Between physical supply restoration and a reconciled remote ledger lies a sequence of hardware boot, name resolution, cryptographic handshake, transaction verification and queue-drain delays.[S13] [S23] [S24] Exhibit 7 illustrates why usable remote synchronisation lags the physical return of mains voltage.[S06] [S23] [S24]
ext4 journals and start systemd.service units.[S21] [S22] [S29]2.12
Warning
Unverified path independence and unrecorded tail risk. The most common survey failure is accepting a vendor's verbal assurance that a backup link is "independent" without walking the cable duct and checking whether the backup base station shares the factory's utility substation.[S03] [S05] Until path_unverified is cleared by a physical route audit and joint outage log, the design must assume that external backhaul can fail simultaneously with utility mains.[S03] [S05]
Read-out
| Mains voltage, phase and frequency events | V, Hz, ms | Three-phase power quality logger at LV board | Continuous event trigger plus 1 min summary | recorder_unpowered | Site electrical engineer |
|---|---|---|---|---|---|
| Generator and UPS transfer interval | ms, s | Digital sub-panel logger and UPS telemetry | Every changeover event plus weekly test | transfer_uninstrumented | Maintenance lead |
| End-to-end WAN commit probe (DNS, TLS, HTTP) | ms, bytes/s, status | Persistent edge probe to test ingestion endpoint | Operator-chosen (e.g. every 60 s in shift) | observer_offline | IT and OT integrator |
| Joint power-and-WAN state occupancy | shift-hours | Joined power and probe ledger (UTC / UUIDv7) | Compiled daily and at baseline close | clock_unsynchronised | Plant operations manager |
| Enclosure ambient temperature and dust | deg C, RH percent | Calibrated data logger inside edge cabinet | Operator-chosen (e.g. every 10 min) | sensor_absent | Reliability engineer |
| Carrier path and tower power dependency | route audit record | Physical duct walkdown and ISP questionnaire | Pre-contract and after any route change | path_unverified | Procurement lead |
Illustrative site-survey dossier template derived from NIST SP 800-82 Rev. 3 (S03), NIST SP 800-34 Rev. 1 (S05) and World Bank survey boundary discipline (S02). Cadences are operator-selected planning inputs, not observed GFT results.
Unit: site baseline specification. Status: illustrative. Sources: S02, S03, S05. Illustrative figures are teaching examples, not measured factory results.
As Exhibit 8 emphasises, the fifth column (Missing-data flag) is as important as the measurement itself.[S02] [S05] Whenever a logger is absent, loses power during a long fault, or lacks a synchronised clock, the dossier records the explicit flag (recorder_unpowered, observer_offline, clock_unsynchronised or path_unverified) rather than substituting a national average or vendor brochure figure.[S02] [S10] [S22]
With the site survey protocol, joint-state matrix and recovery timestamps defined, the next step is to translate the measured interruption envelope into a binding operational contract.[S03] [S05] Chapter 3 establishes how local service classes, recovery-time objectives, recovery-point budgets and explicit degraded operating states turn site observations into testable engineering requirements before any hardware or cloud platform is purchased.[S03] [S05] [S06]
Check the record
Every knot is a source this chapter cites; select one to read its record, or filter the register by grade. Publication dates and inherited-source limits stay explicit. Read the paper's full evidence summary.
30 sources · 279 citation marks on this pageMost cited: S05, 61 times
Grade AGrade BGrade CKnot size: citations on this page
All 30 records shown
R06 · Grade C · inherited record
Inherited evidence R06
About 95 percent of cotton fibre exported raw against 5 percent processed locally; industrialisation target of 50 percent local processing.
Publication date unrecorded. Inherited source-text record. The original publication has not been independently reopened for this web edition.
R07 · Grade C · inherited record
Inherited evidence R07
SND30 restates the 50 percent local-processing target by 2030.
Publication date unrecorded. Inherited source-text record. The original publication has not been independently reopened for this web edition.
R09 · Grade C · inherited record
Inherited evidence R09
Camtext SA integrated cotton-to-textile project, US$300 million (180 billion XAF); Panafritex and Arise with SODECOTON, CNPS and Marlo; 12 November 2025.
Publication date unrecorded. Inherited source-text record. The original publication has not been independently reopened for this web edition.
R10 · Grade C · inherited record
Inherited evidence R10
Camtext SA: 12,000 t a year processed; about 12,000 jobs at the Dibamba hub and 3,000 at Garoua; return on investment near 11 years; optimal capacity around 2033.
Publication date unrecorded. Inherited source-text record. The original publication has not been independently reopened for this web edition.
R12 · Grade C · inherited record
Inherited evidence R12
CICAM capacity utilisation, July 2023 report: Garoua 33.7 percent, CICAM 1 48.9 percent, CICAM 2 10.83 percent; underutilisation cost about CFAF 2 billion in 2020.
Publication date unrecorded. Inherited source-text record. The original publication has not been independently reopened for this web edition.
c02-S01 · Grade A
Enterprise Surveys: Cameroon 2024 Country Profile
All indicators are based on the responses of firms.
2025. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S02 · Grade A
Enterprise Surveys methodology
For these panel firms, sampling weights are adjusted to consider the resulting altered probabilities of inclusion in the sample frame.
2026-10-30. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S03 · Grade A
SP 800-82 Rev. 3: Guide to Operational Technology (OT) Security
This document provides guidance on how to secure operational technology (OT) while addressing their unique performance, reliability, and safety requirements.
2023. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S05 · Grade A
SP 800-34 Rev. 1: Contingency Planning Guide for Federal Information Systems
This document provides guidance to help personnel evaluate information systems and operations to determine contingency planning requirements and priorities.
2010. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S06 · Grade A
SP 800-184: Guide for Cybersecurity Event Recovery
This publication provides tactical and strategic guidance regarding the planning, playbook developing, testing, and improvement of recovery planning.
2016. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S08 · Grade A
RFC 4838: Delay-Tolerant Networking Architecture
The bundle layer forms an overlay that employs persistent storage to help combat network interruption.
2007. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S10 · Grade A
RFC 5905: Network Time Protocol Version 4: Protocol and Algorithms Specification
A client synchronizes to one or more upstream servers, but does not provide synchronization to dependent clients.
2010. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S11 · Grade A
RFC 8633: Network Time Protocol Best Current Practices
It includes recommendations for the stable, accurate, and secure operation of NTP infrastructure.
2019. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S12 · Grade A
RFC 9110: HTTP Semantics
Of the request methods defined by this specification, PUT, DELETE, and safe request methods are idempotent.
2022. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S13 · Grade B
Atomic Commit In SQLite
Atomic commit means that either all database changes within a single transaction occur or none of them occur.
2026-10-30. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S14 · Grade B
Write-Ahead Logging
All processes using a database must be on the same host computer; WAL does not work over a network filesystem.
2026-10-30. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S16 · Grade B
PRAGMA Statements
A transaction committed in WAL mode with synchronous=NORMAL might roll back following a power loss or system crash.
2026-10-30. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S18 · Grade A
RFC 7228: Terminology for Constrained-Node Networks
Such a network may itself exhibit constraints, e.g., with unreliable or lossy channels, limited and unpredictable bandwidth, and a highly dynamic topology.
2014. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S21 · Grade B
systemd.service
The service specific configuration options are configured in the [Service] section.
2026-10-30. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S22 · Grade B
journald.conf
Controls where to store journal data.
2026-10-30. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S23 · Grade A
RFC 9846: The Transport Layer Security (TLS) Protocol Version 1.3
This document specifies version 1.3 of the Transport Layer Security (TLS) protocol.
2026. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S24 · Grade A
RFC 9293: Transmission Control Protocol (TCP)
TCP provides a reliable, in-order, byte-stream service to applications.
2022. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S29 · Grade B
ext4 General Information
Write barriers enforce proper on-disk ordering of journal commits, making volatile disk write caches safe to use, at some performance penalty.
2026-10-30. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S30 · Grade B
Frequently Asked Questions
The rtcfile directive with the -s option can be used to compensate for the RTC drift.
2026-10-30. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S31 · Grade A
RFC 9562: Universally Unique IDentifiers (UUIDs)
A UUID is 128 bits long and requires no central registration process.
2024. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S32 · Grade A
RFC 3339: Date and Time on the Internet: Timestamps
All times expressed have a stated relationship (offset) to Coordinated Universal Time (UTC).
2002. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S0201 · Grade A
RFC 9520: Negative Caching of DNS Resolution Failures
This document updates RFC 2308 to require negative caching for DNS resolution failures.
2023-12. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S0202 · Grade A
RFC 6349: Framework for TCP Throughput Testing
This framework describes a practical methodology for measuring end-to-end TCP Throughput in a managed IP network.
2011-08. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S0203 · Grade A
RFC 2330: Framework for IP Performance Metrics
Any such summary must include the sample size N, because any percentile difference finer than 1/N is below the resolution of the sample.
1998-05. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗c02-S0204 · Grade A
RFC 9499: DNS Terminology
The Domain Name System (DNS) is a simple query-response protocol whose messages in both directions have the same format.
2024-03. Register access date: 2026-10-30. Source provenance is supplied by the paper evidence register.
Open the recorded source ↗S01 World Bank. 2025.
S05 NIST. 2010.
S03 NIST. 2023.
S02 World Bank. 2026-10-30.
R06 Cameroon Trade Hub. Publication date unrecorded; inherited source-text check.
R07 Ecofin Agency. Publication date unrecorded; inherited source-text check.
R12 Business in Cameroon. Publication date unrecorded; inherited source-text check.
S13 SQLite. 2026-10-30.
S31 IETF / RFC Editor. 2024.
S32 IETF / RFC Editor. 2002.
S10 IETF / RFC Editor. 2010.
S22 systemd project. 2026-10-30.
S16 SQLite. 2026-10-30.
S29 Linux kernel project. 2026-10-30.
S18 IETF / RFC Editor. 2014.
S24 IETF / RFC Editor. 2022.
S12 IETF / RFC Editor. 2022.
S23 IETF / RFC Editor. 2026.
S0204 IETF / RFC Editor. 2024-03.
S0201 IETF / RFC Editor. 2023-12.
S0202 IETF / RFC Editor. 2011-08.
S0203 IETF / RFC Editor. 1998-05.
S11 IETF / RFC Editor. 2019.
S30 chrony project. 2026-10-30.
S08 IETF / RFC Editor. 2007.
S14 SQLite. 2026-10-30.
S06 NIST. 2016.
S21 systemd project. 2026-10-30.
R09 Ecofin Agency. Publication date unrecorded; inherited source-text check.
R10 Ecofin Agency. Publication date unrecorded; inherited source-text check.
Keep the thread · chapter 02 of 14
Move the shuttle along the rack to preview any chapter: its sections, exhibits and reading time. Draft web edition. This research is not investment, legal or tax advice; illustrative models are not measured factory results.
The paper's rack · 14 boltsBolt length: reading time
01Part I: Specify the interruption
GFT argues that continuity means retaining a safe operating state and an accountable record through interruption and recovery.
02Part I: Specify the interruption
A useful design envelope records the duration, coincidence and recovery of power and network failures at the intended site.
03Part I: Specify the interruption
Criticality, allowable staleness and restart dependencies should be assigned before choosing equipment or a cloud platform.
04Part II: Keep the local record alive
Protecting a selected record-keeping load can be practical even when the production machinery must stop.
05Part II: Keep the local record alive
Local name resolution, switching, identity and observability need their own continuity plan when external connectivity disappears.
06Part II: Keep the local record alive
The transaction boundary, storage path and acknowledgement semantics decide whether a production record survives abrupt loss.
07Part II: Keep the local record alive
Offline retention and recovery capacity must be designed together so a restored connection can clear its accumulated work.
08Part III: Recover a trustworthy operation
A trustworthy record separates identifiers, source sequence and clock quality so reconciliation can explain what happened during disconnection.
09Part III: Recover a trustworthy operation
An edge model supports continuity only when its inputs, licence, operator interface and fallback remain usable without the cloud.
10Part III: Recover a trustworthy operation
Local credentials, signed software, controlled access and recoverable evidence must keep working through communication loss.
11Part III: Recover a trustworthy operation
Boot order, consistent backups, authorised process release and ledger reconciliation decide whether restoration is complete.
12Part IV: Prove and govern continuity
Compare service classes using transparent operating assumptions and procurement obligations rather than promised resilience percentages.
13Part IV: Prove and govern continuity
A repeatable fault programme should verify declared losses, restart behaviour and reconciled records before a pilot is expanded.
14Part IV: Prove and govern continuity
GFT recommends a staged programme that measures site conditions, proves a local record and expands only when recovery evidence supports it.