Digital Engineering Architecture

Distributed TUFLOW Simulation Scheduling

A shared-storage scheduling approach for coordinating TUFLOW model runs across workstations. This case study describes the queue, process-tracking and monitoring design, with an interactive dashboard prototype.

Standard-user operation Shared-storage coordination Configurable execution controls

Interactive Dashboard Prototype

TUFLOW Administration Dashboard (Simulated Sandbox)
Demo NAS status
Active Node: Station_A | Handshake Recency: 2.1s
Worker CPU
84%
NVidia GPU
92%
Disk Free
1.2 TB
ID Target Station Simulation File Path Execution State Submitter
Station_A C:\TUFLOW_Models\runs\run_M01_EX.bat Running ENG1
Station_A C:\TUFLOW_Models\runs\run_M02_PR.ps1 -d Queue ENG1
Station_B C:\TUFLOW_Models\runs\run_M03_OPT.bat Finished ENG2
đź’ˇ Browser demonstration: task rows, resource readings and status messages are sample data. Buttons preview dashboard interactions; this page does not connect to a NAS or launch TUFLOW.

Scheduling Design Considerations

Permissions & Execution Controls

The design targets operation under standard user permissions, subject to the host organisation’s execution policies. The example configuration includes AllowedScriptPaths, script-signature checks and directory-permission checks. These controls need to be reviewed and tested in the intended environment.

Shared Storage Concurrency Control

Concurrent edits to a shared task register can overwrite updates. The design uses optimistic locking and a stale-lock timeout to manage competing updates and interrupted workers. Recovery behaviour should be checked against the file-sharing service and workstation failure scenarios.

Decoupled Sub-Process Traversal

A launcher can finish before its simulation process exits. The design tracks Windows process identifiers and uses timeout windows to reconcile launcher and child-process states. Detached processes, restarts and interrupted runs are important validation cases.

Hardware Resource Observability

The monitoring design brings together CPU load, GPU utilisation and available disk space to help engineers review workstation capacity. The prototype displays sample values; deployed readings depend on the available hardware and monitoring tools.

Decoupled Multi-Tier System Topology

PRESENTATION

Decentralized Client Dashboards

The proposed dashboard reads shared task state and presents queues and workstation status to engineers.

ORCHESTRATION

Background Daemon Event Loop

The worker loop is designed to check permitted paths, coordinate task updates and track simulation processes. Signature validation depends on configuration and the host environment.

DATA & AUDIT

Shared NAS Registry & Telemetry

A shared task register records queue and execution state alongside workstation telemetry. Audit-log collection and integration with organisational monitoring systems are deployment considerations.

Example Worker Configuration

// Illustrative settings; validate paths, permissions and timeout behaviour before deployment { "NasFolder": "C:\\TUFLOW_Daemon\\Pending_Jobs\\", "AllowedExecutables": [ "TUFLOW_iSP_w64.exe", "TUFLOW_iDP_w64.exe", "powershell.exe" ], "AllowedScriptPaths": [ "C:\\TUFLOW_Models\\" ], "RequireScriptSignature": true, "VerifyAclOnScriptPaths": true, "MinWaitSeconds": 10, "GracePeriodSeconds": 5, "StaleLockThresholdMinutes": 2 }