WDS Scheduling Engine

A Next-Generation Non-Preemptive Scheduler for High-Performance Computing.

Run Demo Custom Input Documentation

Why Do We Need WDS?

Standard Non-Preemptive algorithms have critical flaws in modern workloads.

FCFS (First Come First Serve)

The simplest algorithm. Jobs are executed in arrival order.

  • Suffers from Convoy Effect.
  • One long job blocks all short jobs.
  • Terrible response time for interactive tasks.
SJF (Shortest Job First)

Executes the job with the smallest burst time first.

  • Suffers from Starvation.
  • Long jobs may never run if short jobs keep coming.
  • Ignores User Priority completely.
WDS (Weighted Dynamic Score)

Our proposed hybrid solution.

  • Cooperative Multitasking: Tasks yield to let urgent jobs run.
  • Dynamic Scoring: Balances Speed (SJF) and Fairness (FCFS).
  • Real-Time Capable: Critical Priority cuts the line instantly.
The Core Algorithm

At every time step, WDS calculates a Score for every waiting process. The process with the highest score is selected for the CPU.

Score = (W_wait × WaitTime) + (W_burst × 1/BurstTime) + (W_prio × Priority)
W_wait (Aging Factor)
Prevents Starvation by boosting old jobs.
W_burst (Efficiency)
Mimics SJF by boosting short jobs.
W_prio (Urgency)
Allows VIP tasks to cut the line.

Real-World Applications

Click a card below to see the WDS implementation details.

AI Model Training Clusters

Solving Checkpoint latency in Distributed Deep Learning.
5G/6G Edge Gateways

Critical QoS handling for IoT Emergency Health Alerts.
Autonomous Vehicle OS

Real-time switching for Lidar Obstacle Detection.
Serverless Cloud Functions

Solving the "Cold Start" problem in AWS Lambda/Azure.