A five-second SCADA timeout on a remote platform is not an inconvenience; it can trigger automated safety shutdowns and halt production at a cost measured in thousands of pounds per minute. I have spent over 15 years engineering networks for environments like this, and the pattern is consistent: the sites with the most at stake are usually the ones leaning on a single satellite link that fails in heavy rain, or a cellular connection that saturates every evening. This is how we engineer past that.
Why remote energy sites break networks
An offshore platform or remote wellhead stacks every difficulty at once. Geostationary satellite is often the only day-one option, and its latency, frequently over 600ms, is hostile to real-time SCADA, where a brief interruption can trip a safety shutdown. Rigs are massive steel structures that behave like Faraday cages; sea spray and rain degrade satellite bands; and the atmosphere itself may be explosive, which rules out standard networking gear entirely.
The underlying design fault is nearly always single-link dependency. A satellite link with a strong SLA still suffers weather fade and jitter; edge cellular that runs clean at midday congests by evening. My working definition of resilience in this sector: the physical infrastructure can fail repeatedly without the application ever losing its logical session. Anything less is a backup, not resilience.
Beyond failover: one logical connection
Traditional failover detects a dead primary and switches, and in that gap sessions time out; for SCADA that gap is the whole problem. SpeedFusion takes a different approach: every available transport, VSAT, 5G, 4G, even local wireless, becomes one pool of capacity inside a persistent encrypted tunnel. Lose a physical link and the tunnel stays up on the remaining paths; the monitoring session never re-authenticates, the telemetry never stops.
Two tools matter most for industrial traffic. WAN Smoothing duplicates critical packets across links so the receiving end uses whichever copy arrives first, which is what tames satellite jitter for protocol traffic that cannot tolerate loss. And latency-based routing lets us set hard thresholds, excluding any path that drifts past, say, 150ms for time-sensitive traffic, so a degrading link removes itself from the equation before it hurts the application. For high-throughput work such as live video or remote ROV operation, bonding the combined capacity of cellular and satellite is what makes the workload possible at all.
Hardware for hazardous locations
In this sector, hardware selection is a compliance exercise before it is a technical one. A compressor station or rig doghouse is not the place for an enterprise router. For hazardous-atmosphere zones, Class 1 Division 2 is frequently a hard prerequisite; as one concrete example, the MAX BR1 Mini achieves its C1D2 rating when deployed with Peplink's dedicated mounting kit (ACW-749), and the right certification route varies by model and hardware revision, so we verify against the current datasheet for every deployment rather than assume.
Beyond the atmosphere rating, we look at operating temperature (the MAX series runs from -40°C to +65°C, which matters in uncooled cabinets), shock and vibration standards for units mounted to machinery or transport skids, ingress protection against salt spray and dust, and carrier diversity, with dual or quad-modem models keeping multiple operators live simultaneously. Through steel bulkheads, external high-gain antennas are not optional; the standard paddles will not reach.
Legacy kit is the quiet complication. Plenty of sites still run PLCs and sensors on RS-232 or RS-485, and replacing them is disruptive and expensive. Peplink hardware can bridge those serial protocols into the IP network and carry them over the encrypted tunnel, which brings decades-old assets into centralised monitoring without a rip-and-replace.
Securing the edge
Operational data crossing public cellular and shared satellite backhaul needs security in the transport layer, not bolted on afterwards. SpeedFusion tunnels carry AES 256-bit encryption as standard, and on top of that we apply zero-trust segmentation: OT traffic strictly isolated from crew Wi-Fi and admin networks by VLANs and firewall policy, so a compromised laptop in the mess cannot reach a controller. Firmware stays on a scheduled audit cycle, with the fleet held on the current stable release rather than whatever shipped in the box.
InControl 2 is the management layer that makes a hundred remote sites operable from one screen: mass configuration, coordinated firmware rollouts, GPS correlation for mobile and offshore assets. Where rig managers need something simpler than raw network telemetry, we build portals that translate link health, location and consumption into a view operations staff can actually use.
Scoping a deployment properly
None of this is plug-and-play. A system scoped for basic telemetry will fail the day someone points a high-definition camera feed or remote drilling control at it, so we start from the application requirements, throughput, latency tolerance, protocol behaviour, then design backwards through carrier coverage, environmental risk and failure points. Enabling bonding is trivial; tuning sub-tunnel parameters to handle VSAT jitter and congested 4G loss at the same time is where deployments succeed or fail, and it is where the 15 years earn their keep. We train your team to run the environment day to day, and our managed service backs them for the faults that would otherwise mean a helicopter.
The short version
Diversify the transport, bond it into one persistent tunnel, hold hard latency thresholds for SCADA, buy hardware certified for the zone it lives in, segment OT from everything else, and manage the fleet centrally. We avoid the marketing hype of "unbreakable" connections and deal in uptime, latency and packet delivery instead. If SCADA timeouts or satellite overages are a live problem on your sites, get in touch for a scoping conversation.
Frequently asked questions
Can SpeedFusion really keep SCADA sessions alive through a link failure?
Yes, that is its design purpose. The session rides a logical tunnel spanning all links, so a physical failure reduces capacity rather than dropping the session. The transition is typically sub-second when the deployment has been sized with headroom.
Is satellite alone ever enough for a remote site?
For basic, loss-tolerant telemetry, sometimes. For real-time control traffic, geostationary latency and weather fade make it a poor sole transport; paired with cellular in a bonded tunnel it becomes a strong contributor instead of a liability.
What certifications should I look for in rig-side hardware?
Class 1 Division 2 for hazardous atmospheres where the zone requires it, appropriate ingress protection for salt and dust, wide-range operating temperature, and recognised shock and vibration standards for machinery mounting. Certification routes vary by model and revision; check the current datasheet, or ask us to.
Can old serial equipment join a modern monitored network?
Usually, yes. RS-232/RS-485 devices can be bridged through the Peplink gateway into the IP network and carried over the encrypted tunnel, giving legacy PLCs and sensors centralised visibility without replacement.
How do you keep crew internet away from operational systems?
Strict segmentation: separate VLANs, firewall policy that denies crossover by default, and OT access limited to what each role genuinely needs. The crew network should be able to fail, or be compromised, without operational traffic noticing.