Oil & Gas
What happens at the rig site should be visible at head office in minutes — not the next morning, and not never.
What’s actually hard here
The constraints that decide the architecture.
Not the sector’s challenges in general. The specific things that make software for it hard, and that anyone proposing to build it should already know.
The link is not a network
The connection to a rig fails regularly, unpredictably, for anything from seconds to a day. Software that treats that as an error condition spends most of its life in an error condition. The link being down has to be a normal operating state — which means the rig-site component is the real system and the cloud is a consumer of it, not the other way round.
A healed gap is a lie
When a link returns and buffered data arrives, a system that seamlessly merges it produces a chart with no seam — and no way to know it was ever blind. On a well, when you were blind is precisely the question asked afterwards. Backfilled periods have to stay marked, even though marking them makes the log look worse.
Depth and time are both real
The same reading belongs at a depth and at a time, and the relationship between the two is not fixed — during a connection, time passes and depth does not. Store one index and derive the other and every log goes wrong exactly where drilling stopped. Carrying both as first-class costs storage and complexity, and is the difference between a driller and a geologist reading the same well without arguing about the chart.
Data custody is an architecture input, not a policy annexe
Who may hold well data, in which country, and what a partner may see — these decide the deployment before any diagram is drawn. The industry's data standards let systems exchange data; they do not decide who is allowed to keep it. Anyone proposing an architecture before asking those questions is proposing their architecture.
How we engineer for it
The practices this sector actually needs.
Relevant platform
We already built this.
Because we kept meeting the problem. It is not the only thing we can do for this sector — it is the part we had solved enough times to engineer properly.
Tell us what cannot go wrong.
If the constraints above look like your constraints, that is the conversation we want. You will speak to an engineer.