Process historians, time-series storage, sampling, trending, compression, and retention. 89 terms in this topic - each defined in plain English on its own page.
Process historians, time-series storage, sampling, trending, compression, and retention.
The family of avg, min, max, count, and sum that collapses many samples into one, and which answers which question.
Read → GlossaryTime-bounded archive chunks for retention, backup, and query pruning, and what happens at an archive shift.
Read → GlossaryMap timestamp, value, and quality columns, handle timezone and duplicates, tag as substituted, then verify against archive rules.
Read → GlossaryThe older two-test compression that predates swinging door, and why it leaves stepped, jagged trend artifacts.
Read → GlossaryDerived historian points computed from other tags, and the recomputation problem when their source data is backfilled.
Read → GlossaryCheap deep object-storage tiers with minutes-to-hours retrieval, and how that latency shapes what stays cold.
Read → GlossaryTracking recurring filing deadlines and how long measurement and report records must be kept and stay auditable.
Read → GlossarySingle grab versus flow- or time-proportional composite, and when each is required for custody or compliance.
Read → GlossarySet a per-tag deviation limit from instrument noise and required fidelity, then validate against raw data and reconstruction error.
Read → GlossaryLargest gap between the historian's interpolated line and the real signal, bounding the worst-case error in any compressed value.
Read → GlossaryAn automatically and incrementally maintained pre-aggregated table that makes repeated dashboard queries fast.
Read → GlossaryHow counting samples per bucket exposes data completeness and collector health before value aggregates mislead.
Read → GlossaryThe share of expected samples actually captured over a period - the headline coverage KPI for historian data.
Read → GlossaryHow historians mark and query around genuine missing intervals versus stale flatlines, and why gaps are not interpolated.
Read → GlossaryAutomated aging rules that move data hot to warm to cold to deletion on a schedule as it ages.
Read → GlossaryPre-computed hourly and daily summaries stored beside raw data to speed long-range queries and reports.
Read → GlossaryWhy keeping every Nth sample loses spikes and aliases, and how signal-aware compression avoids it with bounded error.
Read → GlossaryReducing millions of points to a chart's pixel width with min/max or largest-triangle methods without hiding spikes.
Read → GlossaryTwo values for the same tag at the same time; historians overwrite, reject, or keep both, and unresolved duplicates corrupt aggregates.
Read → GlossaryA fixed-size circular store that buffers unsent samples and overwrites the oldest data when full.
Read → GlossaryHow a gateway snaps mismatched sensor readings onto a common time grid so downstream calculations line up.
Read → GlossaryThe mini sparkline beside a value on a process graphic that reveals its recent trajectory at a glance.
Read → GlossaryReports that surface only out-of-limit conditions - management by exception - and how it differs from report-by-exception telemetry.
Read → GlossaryThe two distinct deadband stages in a historian: the exception test at the collector versus compression at the archive.
Read → GlossaryAggregate computed only over the time a condition was true, such as average pressure while the compressor ran.
Read → GlossaryRecording every packet to a rolling PCAP store for forensic replay, and the storage-versus-rewind trade-off.
Read → GlossaryText note attached to a value or time that adds context without changing the value, audited and never compressed away.
Read → GlossaryInserting recovered or buffered data into the archive after the fact, and the ordering and recompression it forces.
Read → GlossaryConfigurable span within which the archive accepts past-timestamped writes; too narrow a window silently drops late data.
Read → GlossaryThe interface that reads tags from field devices and SCADA and streams them into a process historian, with buffering and deadband filtering.
Read → GlossaryShare of scanned values a historian stores after filtering; outliers from the normal high range flag a mis-tuned tag fast.
Read → GlossaryRe-running calculated tags and rollups over a window after backfill or a correction so derived values stay consistent.
Read → GlossaryQuery the historian over ODBC by tag, time, and interval in raw or interpolated mode, and avoid full-archive scans.
Read → GlossaryQuerying many site historians as one without copying all their data into a single central store.
Read → GlossaryThe write side of a historian: sample rate times tag count sets a sustained IOPS load that usually needs SSD or NVMe.
Read → GlossaryCollector writes the same stream to two independent archives so either can serve queries and survive loss of the other.
Read → GlossaryDetects the window one member missed offline, copies the missing archive data back, and verifies both copies match.
Read → GlossaryLookup that maps tags and time ranges to archive blocks so queries seek instead of scanning; when to rebuild it.
Read → GlossaryHow often the collector samples a source before compression, the oversampling tradeoff, and how it differs from PLC scan cycle.
Read → GlossaryEstimating disk needs from tags, sample rate, bytes per value, and retention, then applying compression.
Read → GlossaryTags bundled by collection frequency so a collector polls each group on its own interval, balancing load.
Read → GlossaryArchive-side limit on ingest rate so a burst of collectors flushing buffers cannot overwhelm I/O or stall live data.
Read → GlossaryPlacing a past window like yesterday or a golden run over the live trace to compare against a reference.
Read → GlossaryLossless on-disk encoding - delta-of-delta timestamps, XOR and run-length values - that shrinks storage without dropping points.
Read → GlossaryIntegrating a rate over a window by trapezoidal area to get an accumulated total like volume or energy.
Read → GlossaryMatching sample withdrawal velocity to line velocity so entrained water and sediment are captured in true proportion.
Read → GlossarySparse, hand-entered values like lab assays and tank gauges, and how historians handle their irregular, out-of-order arrival.
Read → GlossaryThe sample-and-hold query mode that returns the most recent stored sample at any timestamp.
Read → GlossaryHow a historian rebuilds a value at any timestamp from the two bounding compressed points.
Read → GlossaryBundling several physical links into one logical LAG for bandwidth and uplink redundancy between managed switches.
Read → GlossaryCollecting logs from PLCs, HMIs, and network gear into one tamper-evident store as the raw material a SIEM analyzes.
Read → GlossaryA decade-scaled trend axis that keeps values spanning many orders of magnitude readable across the whole range.
Read → GlossaryDeadband and swinging-door as bounded-error lossy storage, and how the deviation guarantees a maximum reconstruction error.
Read → GlossaryThe LTTB algorithm picks the most visually significant points so a downsampled chart still looks like the full data.
Read → GlossaryTag with values typed in by a person, timestamped and flagged as manual, with an audit trail of who entered them.
Read → GlossaryKeeping the high and low of each pixel column so a downsampled trend still shows transients averaging would erase.
Read → GlossaryThe subscriber-writer pipeline that decodes MQTT payloads, maps them to measurements, and batches them into a time-series store.
Read → GlossaryA saved chart of related tags plotted together on one time axis to reveal cause, effect, and correlation.
Read → GlossaryZero is a measured none, null is no data; summing gaps as zero understates totals, null-aware aggregates skip them.
Read → GlossaryThe twice-the-highest-frequency rule for sample rates and how undersampling folds fast signals into false slow ones.
Read → GlossaryWhat breaks when points arrive earlier than stored data, from buffering or clock skew, and how historians reorder them.
Read → GlossaryHow P95, P99, and the median expose peak loads and tail behavior that an average smooths flat.
Read → GlossaryDisplay retrieval returning min, max, first, and last per pixel-interval so long trends draw fast without losing spikes.
Read → GlossarySetting how long a historian keeps full-resolution raw samples versus pre-computed aggregate rollups.
Read → GlossaryEvenly spaced fixed-interval logging versus exception-based data at uneven times, and why the difference breaks averaging.
Read → GlossaryThe mechanism that automatically expires and drops data past a set age, configured per measurement as a TTL.
Read → GlossaryResidence time in a separation vessel - why it governs how cleanly oil, water, and gas separate, and how it's sized.
Read → GlossaryHow often a sensor is read versus how often a value is stored, and why the gap saves space but can miss spikes.
Read → GlossaryClearing stale product and dead volume from sample lines before a grab, and how flush cycles are sequenced.
Read → GlossaryThe workflow to take a representative sample: purge the point, choose a container, label, and keep chain of custody.
Read → GlossaryRecurring automated job that writes a tag set to a file or endpoint on a schedule for downstream feeds and handoffs.
Read → GlossaryThe count of unique tag and label combinations that drives memory and index cost and can explode with careless labeling.
Read → GlossaryThe average time biomass stays in an activated-sludge system, set by wasting rate and governing nitrification.
Read → GlossaryHow a per-bucket stddev quantifies process variability and noise for control-loop health and SPC.
Read → GlossaryStep through the SDT algorithm: how the upper and lower doors pivot and collapse to trigger a stored point.
Read → GlossaryStoring recent data hot, mid-age data warm, and old data cold to balance query speed against cost.
Read → GlossaryHow a historian query groups raw samples into fixed intervals with min, max, avg, and sum functions.
Read → GlossaryWhy compressed data must be averaged by area under the curve, not as a plain mean of stored points.
Read → GlossaryHow historians integrate a rate into a running total and handle rollover, resets, and daily-boundary flow totals.
Read → GlossaryAn operator note pinned to a point on a trend so the reason behind a change travels with the data.
Read → GlossaryA movable line that reads exact values and times off a trend, with dual cursors measuring change and rate.
Read → GlossaryDetecting when a value drifts from its expected trend or baseline - catching slow degradation that fixed alarms miss.
Read → GlossaryA how-to on defining reusable named trend groups: choosing tags, setting scales and axes, and a time span.
Read → GlossaryPlotting a tag's live values over its own past run on shared axes to spot drift and abnormal behavior.
Read → GlossaryEach plotted line on a SCADA trend, named for the ink pens of chart recorders.
Read → GlossaryWhy storing UTC and displaying local avoids the DST duplicate and missing hour across multi-timezone sites.
Read → GlossaryThe master list of every SCADA data point - tags, types, metadata, and why the tag database is the heart of any SCADA configuration.
Read → GlossaryThe database built for timestamped sensor data - compression, time-range queries, TSDB vs relational, and its role in SCADA historians.
Read → GlossaryHow historians drop insignificant changes to store less - simple deadband vs swinging door, and choosing compression without losing real events.
Read →Merobix is cloud SCADA for oil & gas - it reads the instruments and protocols defined here into one dashboard.
We respond within 1 business day.