How does Vardar connect without adding inline control risk?

Vardar begins with a read-only copy of traffic from SPAN or TAP and can add customer-approved read-only evidence sources. Active integrations remain explicitly governed by customer policy and approval.

Does raw OT data leave the site?

The Vardar Edge AI appliance performs decoding, correlation, inference and investigation locally. Raw OT traffic and process values remain inside the customer environment.

How does Vardar reduce false positives?

Vardar compares two evidence paths: network behavior and process or physics behavior. Agreement raises confidence; a single evidence path stays under investigation instead of being presented as confirmed cyber.

Attack #20, AIT504 (SWaT 2015)

1.0Plant

Where it hit the plant

raw water intake

chemical dosing

ultrafiltration

dechlorination / UV

reverse osmosis

AIT504

permeate transfer & backwash

2.0Timeline

Attack window and detections

Labelled window2015-12-29 18:15:43 to 2015-12-29 18:22:17plant local time

Physical Residual

instant

Temporal Sequence

instant

Process Invariants

no flag

Fused Ensemble

instant

Shaded band: the labelled attack window from the iTrust ground-truth catalog. Marks: the first second each engine flagged the window. "No flag" means the engine stayed silent for the whole window.

3.0Evidence

Why it was flagged

At 2015-12-29 18:15:43, attack #20 (SSSP) manipulated AIT504 — “Set value of AIT-504 to 255 uS/cm”. The serving model raised the alert immediately — within the first second of the attack (peak score 8.3845x the operating threshold). The dominant physical evidence: AIT504 — RO permeate conductivity analyzer, Stage 5 (reverse osmosis) ran 81.3 standard deviations above the value predicted by the plant's learned physics, carrying 25% of the anomaly score — 15.2x its share in normal operation. Supporting deviations were concentrated in AIT503, P403, AIT201. In plain terms: the plant's own physics stopped agreeing with what the sensors were reporting, in exactly the part of the process the attacker touched.
From the official explain run 2026-06-12T06:42:56Z

Single channelPeak score 8.38x of the operating thresholdFused peak 1.00x of the operating threshold

Sensors carrying the anomaly score

Sensor	Share of score	Deviation	Share vs normal
AIT504 RO permeate conductivity analyzer, Stage 5 (reverse osmosis)	25.3%	81.3σ	15.2x
AIT503 RO feed conductivity analyzer, Stage 5 (reverse osmosis)	16.0%	51.5σ	4.3x
P403 UV/RO feed pump (standby), Stage 4 (dechlorination / UV)	14.9%	47.8σ	14.6x
AIT201 conductivity analyzer (NaCl dosing control), Stage 2 (chemical dosing)	8.2%	26.2σ	1.8x
AIT501 RO feed pH analyzer, Stage 5 (reverse osmosis)	5.1%	16.3σ	1.4x
PIT501 RO feed pressure transmitter, Stage 5 (reverse osmosis)	3.5%	11.4σ	1.6x
FIT501 RO feed flow meter, Stage 5 (reverse osmosis)	3.0%	9.6σ	2.1x
LIT301 UF feed tank level transmitter, Stage 3 (ultrafiltration)	2.7%	8.8σ	0.5x

4.0Provenance

Every figure has a run.

Validation run

2026-06-11T22:34:18Z, commit 97752838a3f3d5271b7e43335fcf1f839fdf057a

Model image

sha-97752838a3f3

Metrics artefact SHA-256

80c98149b000d88af3fc7b58ab407a24c81f3839426e7d4e53c2011bbf15c52a

Explain run

2026-06-12T06:42:56Z

Explain artefact SHA-256

2616f67199c281fc558b4b05e3cccdeb223910e4c07addf0c80bc50f0ec5b333

Generator

vardar-portal/scripts/build-gallery-data.mjs (DEL-34/DEL-43)

SWaT and WaDi datasets courtesy of iTrust, Centre for Research in Cyber Security, Singapore University of Technology and Design. Detection figures trace to the reproducible model runs shown with each dataset.

More caught attacks from SWaT 2015Attack #28 Attack #22 Attack #37 Attack #6 All caught attacks

The next record is your process.

A technical evaluation on one critical process, with success criteria agreed in advance, produces the only figure that applies to your site.

Evaluate one critical process

Attack #20AIT504

Sensors carrying the anomaly score

The next record is your process.