ML DDoS Detection and Mitigation
Budget: ₹1,500 – ₹12,500 INR
I need a complete workflow that can automatically spot a DDoS attack against my containerised workload, show the live impact in an intuitive dashboard, and then take action by moving the affected container to a clean node.
Here is how I picture the flow:
• Build and train a Random Forest model that learns to recognise DDoS patterns from the data I will provide.
• Expose the model as a real-time detector that watches the cluster continuously.
• Stream the model’s alerts and key metrics into Grafana (or a comparable tool you recommend) so I can see traffic spikes, prediction scores and mitigation status at a glance.
• When an attack is confirmed, trigger an automated response that both throttles or blocks the malicious traffic and initiates the container migration to a predefined “safe” node within the same cluster, ensuring minimal downtime.
Deliverables must include:
1. Clean, well-commented source code for the Random Forest pipeline and the alerting logic.
2. A working Grafana (or alternative) dashboard export that I can import and run immediately.
3. Scripts or Kubernetes manifest(s) that handle the mitigation steps and the container migration.
4. A brief README explaining setup, configuration variables and how to extend the model with new data.
I will validate the solution by launching a controlled DDoS test; the model should detect it, the dashboard should visualise it in near real time, and the migration should complete without service interruption. Feel free to suggest refinements, but please keep Random Forest as the core algorithm.
Here is how I picture the flow:
• Build and train a Random Forest model that learns to recognise DDoS patterns from the data I will provide.
• Expose the model as a real-time detector that watches the cluster continuously.
• Stream the model’s alerts and key metrics into Grafana (or a comparable tool you recommend) so I can see traffic spikes, prediction scores and mitigation status at a glance.
• When an attack is confirmed, trigger an automated response that both throttles or blocks the malicious traffic and initiates the container migration to a predefined “safe” node within the same cluster, ensuring minimal downtime.
Deliverables must include:
1. Clean, well-commented source code for the Random Forest pipeline and the alerting logic.
2. A working Grafana (or alternative) dashboard export that I can import and run immediately.
3. Scripts or Kubernetes manifest(s) that handle the mitigation steps and the container migration.
4. A brief README explaining setup, configuration variables and how to extend the model with new data.
I will validate the solution by launching a controlled DDoS test; the model should detect it, the dashboard should visualise it in near real time, and the migration should complete without service interruption. Feel free to suggest refinements, but please keep Random Forest as the core algorithm.