The platform loses satellite in the storm
Exactly when the site is hardest to reach physically, the weather takes the link that monitoring depends on.
Critical infrastructure & utilities
Monitoring that outlives the link.
For utility and grid operations, offshore oil and gas platforms, energy operators, pipeline and windpark asset managers: control and telemetry that keep flowing when a carrier fails, the weather fades the dish, or the backhaul is cut.
The stakes
The stakes
A nominally redundant design can still share one provider, one mast, one trench, one upstream.
SCADA goes dark mid-fault
The moment operators most need telemetry is the moment the saturated or failing path drops it.
A telemetry gap becomes a compliance incident
Regulated monitoring doesn’t excuse outages; the historian’s missing hours have consequences beyond operations.
One connection over everything
Dissimilar paths, one observable continuity layer.
Nexus Atlas turns the site’s mixed inventory — satellite, dual cellular, microwave, private WAN, industrial radio — into one encrypted connection with per-class delivery, and exposes whether the inventory provides genuine failure-domain diversity.
VSATdual cellularmicrowaveprivate WANindustrial radio
Three pillars
Built for unattended sites.
Each pillar is a demonstrable behavior against the site’s real failure domains.
Predict the fade
Signal-trend projection moves traffic off a degrading path while it still carries — rain fade becomes a reweighting event, not an outage.
Control traffic first
SCADA, alarms and safety telemetry ride the priority classes; bulk logging and maintenance data yield when throughput contracts sharply.
No telemetry gaps
Buffered delivery on reconnection: through a blackout, telemetry waits in store-carry-forward and the historian receives a complete series.
0.1–0.75 sTypically configured reweighting cadence as path quality changes.
0.0008%Effective loss from three independent links each losing 2% (broadcast mode).
6 classesDelivery contracts — control and alarms always first.
100% softwareOn the industrial computers the site already runs.
VSAT + LTE + microwave through the weather.
As the satellite path fades, traffic reweights toward coastal cellular and microwave; when the sky clears, capacity rebalances — with no operator action at either transition.
Dissimilar bearers, genuinely diverse.
Dual cellular from different carriers plus a private WAN path — and Atlas’s measurement makes visible whether that diversity is real or shares one upstream.
Mark a site camera or a live telemetry stream for retention and the monitoring record is complete after an outage: the live view drops what arrives late, but the site keeps every packet of the stream, the control room asks for exactly what it missed, and only those packets cross — on the lowest lane, behind alarms and control — so the record kept for review has every packet once, in order. How the gaps are filled
Admitted by certificate, with no entry on any hub.
Sensors and gateways leave the factory with a certificate from your own fleet authority and are admitted wherever that authority is trusted — no per-node entry on any hub, nothing edited when a site adds a device, and every admission recorded with who made it. A removed device is refused everywhere from the next signed list. Join by certificate
A release that cannot strand a substation.
Keys and the shared secret rotate on a signed schedule with windows set per fleet, and designated lifeline nodes keep answering a node’s old material for one more window — so a site visited once a year, out of contact across a rotation, comes back on its old key, takes the new one over its first connection and switches by itself; one away past the end rejoins the way a new node does. A configuration release never completes while a required substation’s node is unconfirmed: silent at the deadline, the change withdraws itself, the previous settings go back out signed, and the operator is told which site held it. Keys without a flag day The release gate
A maintenance link that survives the environment.
Robotic maintenance platforms reach operators through relayed paths across the park’s own nodes — continuity framed as capability, not as a named customer’s deployment.
A device that leaves its site isolates itself.
Each node can carry a signed operating area. A gateway lifted from its cabinet or a sensor driven off the site refuses every peer and tears its sessions down once it has been outside beyond its grace time, and lifts that by itself when it is back inside; where policy asks, it wipes its keys instead. A node without a position fix never acts, because a jammed receiver proves nothing. The signed operating area
The contractor’s network meets yours at one gateway.
A maintenance contractor runs its own network across the site. One gateway is a member of both and bridges them: the operator’s control room reaches the contractor’s chosen sensors at addresses of its own network, and no field device is reconfigured. Each side decides which hosts take part, and the contractor can revoke the gateway like any member. Opt-in. How the seam works
A contractor who must only watch gets a link, not a seat on the network: the feed gateway — a separate package beside Nexus Atlas — sends one site’s camera outward from a read-only member to a feed server; the contractor opens it in a browser with a token for that one feed, until it expires or is cut with one click, and the player draws the viewer’s name across the picture. The site opens no port, and nothing travels inward. How the gateway works
Pure software. Pilot on the site computers you already operate — no procurement. The data plane runs with no cloud dependency; the optional hosted console adds fleet alerts and audit where policy allows it. Deployment & operations.