Dispatch systems built to run in production.
Real-time fleet tracking, draw-point control, production reporting. Laravel platforms running on underground operations since 2013. Project-quoted.
What dispatch systems actually do
A dispatch system is the control layer between the mine plan and the equipment carrying it out. Underground that means real-time tracking of LHD fleets across active levels, RFID positioning through the tunnels, draw-point order publication, hydraulic and tilt sensor telemetry off vehicle cabs, and shift-by-shift tonnage reporting against plan. The unglamorous parts are what make the platform credible: MQTT brokers running over patchy underground WiFi, audit-grade logging that the inspectorate actually reads, dual-database write paths that survive a network drop mid-shift, and custom HMI hardware in vehicle cabs running operator interfaces no off-the-shelf product ships with.
Engagements
Palabora Mining Company
Block-cave operation. BCD Dispatch coordinates LHD fleets across drawpoints in real time, tracking production against draw card targets to protect ore recovery — uneven draw costs 15-25% of total ore. Laravel 12 + Livewire 3 + PostgreSQL + MQTT over SSL underground. Six user roles on one platform — dispatchers, supervisors, planners, operators, maintenance, management — 24/7 in production. Built equal-partner with Graybeard Solutions (hardware integration, MQTT broker, IFM HMI panels in cabs).
Cullinan Diamond Mine
Underground diamond mine, four active mining levels, running our dispatch and draw-point control platform for roughly ten years. Real-time LHD fleet tracking through an RFID positioning system threaded across the tunnels, MQTT broker bridging surface and underground over the site WiFi, custom IFM HMI touchscreens in vehicle cabs, and hydraulic-sensor telemetry (bucket tilt, boom lift) feeding load-tracking and production monitoring. Laravel and Livewire on IIS against a SQL Server dual-database backend, PWA-capable for the operator layer, three shifts a day, tonnage targets reconciled per shift and per draw point. 24/7/365 since cutover.
How we approach dispatch projects
-
1. Discovery
Two to three weeks on average. At least one site visit — underground, on the active levels. Integration audit of the existing ERP and comms stack. Stakeholder workshops that put mine planning, operations, and IT in the same room before any architecture decisions are made.
-
2. Architecture
Data flow gets sketched end to end before any code is written: vehicle telemetry to on-board HMI to underground or surface WiFi to MQTT broker to Laravel app to SQL Server or Postgres. Redundancy posture, offline behaviour, and dual-write paths for network drops are decided up front, not patched in later. Stack named with a reason against it.
-
3. Integration
ERP endpoints, fuel-management systems. Custom hardware too — IFM touchscreen HMIs in vehicle cabs, RFID tag readers along tunnel reference points, hydraulic sensors. Most of these vendors pre-date the dispatch system; we adapt to their protocol where the kit can't be modified, and build the bridge ourselves.
-
4. Handover
Operator training is run across all three shifts so nobody learns it second-hand from a colleague. IT team gets a written runbook covering deploy, rollback, log paths, and on-call escalation. Cutover is shift-aware — never mid-shift — and dual-write reliability stays on through the first month so a comms outage doesn't lose production data.
-
5. Maintenance
A separate enterprise retainer, not a Site Care plan — scale is different, SLA expectations are different. Quarterly review of incidents and capacity, named on-call engineers, and patch cycles scheduled against shift changeovers rather than vendor-default Patch Tuesday. Some sites have been running this way with us for the better part of a decade.
Common questions about dispatch projects
What's a typical engagement length?
What stack do you use?
Onsite or remote?
What does post-launch support look like?
Hardware integration — what's in scope?
Around ten years of dispatch system work across South African mining, with two named clients still in production: Cullinan Diamond Mine running underground on the platform for roughly a decade, Palabora Mining Company running at block-cave scale across multiple production years. Audit-grade logging, shift-aware patches, and dual-write reliability are the routine parts — every integration past that is novel, because every site's hardware history is novel. We think each one through and solve it deliberately, and the platforms hold: in production for a decade, maintained by the team that built them, still ours.
Smaller scope? See custom web development →