51-Pipeline Card Template v1.0 | Chapter 5 — Timebase, Synchronization & Buffering (Sampling Consistency)

Home ／ Docs-Technical WhitePaper ／ 51-Pipeline Card Template v1.0

Chapter 5 — Timebase, Synchronization & Buffering (Sampling Consistency)

I. Purpose & Scope

• Standardize timebase, synchronization, and buffering/backpressure models, thresholds, and validation flows to ensure sampling consistency and temporal traceability across batch/streaming/near-RT pipelines.
• For stages involving path quantities (arrival time/phase), the text must explicitly show gamma(ell) and d ell, and the data side records delta_form ∈ {general, factored}; publication requires p_dim = 1.0.

II. Prerequisites & Inputs

• Timebase model: specify f_s (sampling rate), t_exp (exposure/dwell), cadence, and window T_win.
• Sync chain: prefer GNSS 1PPS + 10 MHz; fallback PTP → NTP.
• Metrology & citations: pass I70-dim_check with check_dim_report.json; citations use “volume + version + anchor (P/S/M/I)”, anchor coverage ≥ 90%.
• Path consistency: len(gamma_ell)=len(d_ell)=len(n_eff)≥2, Δell meets sampling constraints.

III. Timebase Modeling

1. Sampling constraints:
   • Nyquist: f_s ≥ 2·f_max; in micro-batch, batch_window = k / f_s.
   • Path step: Δell ≤ ( c_ref / f_s ) / max(n_eff); phase time step: Δt ≤ λ_ref / ( 2 c_ref ).
2. Timestamp semantics: unified ISO-8601; define acq.ts_start/ts_end and timezone; alignment strategy align(ts, T_win).
3. Windows & alignment: sliding/tumbling windows require aligned boundaries and bucketing rules (drop/round/nearest).

IV. Synchronization Chain

• Sources & fallback: GNSS 1PPS (primary) → PTP (preferred) → NTP (last resort).
• Health metrics: δt_abs (absolute offset), Δτ_ch (inter-channel skew), σ_y(τ) (Allan deviation).
• Lock criteria: clock_state="locked" and
  |δt_abs| ≤ τ_abs, |Δτ_ch| ≤ τ_ch, σ_y(τ) ≤ σ_target.
• Loss of lock: enter fallback and elevate uncertainty labels; if unrecoverable, degrade the stage or reject with [Restricted].

V. Buffering & Backpressure

• Buffer strategy: fifo | watermark | time-based | count-based; recommend time watermark + event count dual thresholds.
• Backpressure strategy: drop_tail | shed | slowdown | requeue; prefer slowdown for low-risk, flag high-risk with quality.flags += {"buffer_alert"}.
• Disorder & latency: define out-of-order window T_late and max tolerated latency T_max; exceeding thresholds triggers compensation or rejection.

VI. Sampling Consistency for Path Stages

1. Required inputs: gamma(ell), d ell, n_eff(ell), c_ref, and (for phase) λ_ref; record delta_form.
2. Unified forms (two equivalent):
   • T_arr = ( 1 / c_ref ) * ( ∫ n_eff d ell )
   • T_arr = ( ∫ ( n_eff / c_ref ) d ell )
     Phase: Phi = ( 2π / λ_ref ) * ( ∫ n_eff d ell ).
3. Alignment order: time alignment (window/watermark) → path alignment (gamma_ell/d_ell/n_eff sync) → phase alignment (reference window).

VII. Gates & Thresholds

• G3 | Path conventions: gamma/measure/delta_form present; len(path)≥2.
• G4 | Dimensional closure: check_dim_report.json passes; p_dim = 1.0.
• G5 | Freshness: clock_state="locked"; |ts_start − calib.timestamp| ≤ τ_calib.
• Jitter/offset thresholds: |δt_abs| ≤ τ_abs, |Δτ_ch| ≤ τ_ch, σ_y(τ) ≤ σ_target; breaches trigger degrade/fallback.

VIII. Machine-Readable Configs
A. timebase.yaml

version: "1.0.0"

sampling:

fs_hz: 2000

t_exp_s: 0.01

cadence_s: 0.05

windows:

type: "tumbling" # tumbling|sliding

size_s: 1.0

align: "epoch"

nyquist:

f_max_hz: 800

path_constraints:

delta_ell_max_m: "c_ref/fs_hz/max(n_eff)"

phase_limit:

dt_max_s: "lambda_ref/(2*c_ref)"

B. sync_config.yaml

version: "1.0.0"

sources: ["GNSS_PPS","PTP","NTP"]

thresholds:

abs_ns: 50

skew_ns: 5

allan_1s: 1.0e-11

fallback:

order: ["GNSS_PPS","PTP","NTP"]

on_fail: ["increase_uncertainty","mark_quality_flag","switch_mode"]

C. buffer_policy.yaml

version: "1.0.0"

buffer:

strategy: "watermark"

capacity: 4096

lateness:

T_late_s: 0.5

T_max_s: 2.0

backpressure:

mode: "slowdown" # slowdown|drop_tail|shed|requeue

shed_threshold: 0.9

alerts:

topics: ["buffer_alert","sync_alert"]

IX. Validation & Monitoring

• /validate output: report G3/G4/G5 and jitter/offset thresholds; failures trigger S1–S5.
• Online KPIs: Latency_P50/P95, Throughput, σ_y(τ), δt_abs, Δτ_ch, buffer_fill_ratio, drop_rate, Q_res, p_dim.
• Alert rules: lock loss / PD failure / gate breaches / path desync; support suppression windows and escalation.

X. Anti-Patterns & Fixes

• Anti: T_arr = ∫ n_eff / c_ref d ell (missing parentheses) → Fix: T_arr = ( ∫ ( n_eff / c_ref ) d ell ).
• Anti: declaring only gamma(ell) without d ell/delta_form → Fix: add both and align n_eff length.
• Anti: accepting writes when clock_state!="locked" → Fix: reject or fallback and tag [Restricted].

XI. Release & Layout

PTN_EXPORT/

configs/

timebase.yaml

sync_config.yaml

buffer_policy.yaml

reports/

check_dim_report.json

validate_report.json

audit.jsonl

figs/

sync_health.pdf

buffer_watermark.svg

path_alignment.pdf

report_manifest.yaml

SIGNATURE.asc

XII. Cross-References

• Architecture & graph: Ch. 3; Stage specs & control equations: Ch. 6; Gates: Ch. 9; UQ coupling & feedback: Ch. 10.
• Parameter/Error/Protocol templates: corresponding chapters in Parameter Card, Error Budget Card, Experimental Protocol Card.

XIII. Checklist

• timebase.yaml / sync_config.yaml / buffer_policy.yaml configured and scenario-appropriate.
• f_s ≥ 2·f_max; Δell ≤ ( c_ref / f_s ) / max(n_eff); Δt ≤ λ_ref / (2 c_ref).
• clock_state="locked"; |δt_abs|/|Δτ_ch|/σ_y(τ) within thresholds; fallback defined.
• Path fields explicit gamma/measure/delta_form; len(path)≥2; alignment order correct.
• /validate passes G3/G4/G5; check_dim_report.json attached, p_dim = 1.0.
• Release bundle includes configs/reports/figures & signature; all citations have anchor coverage ≥ 90%.