Skip to content

mojtabagolab10/TPMCD-Cloud-Container-Cost-Optimization

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 

Repository files navigation

TPMCD: Hybrid Container–VM Optimization Framework for Cost and Throughput in Cloud Datacenters (2025)

Hybrid container–VM scheduling framework for cost, throughput and execution-time optimization in cloud datacenters (TPMCD, 2025).

🔗 DOI

https://doi.org/10.1016/j.jnca.2025.104132

Published in: Journal of Network and Computer Applications (Elsevier)


🌍 Executive Summary

Modern cloud datacenters operate under extreme pressure:

  • Increasing workload volumes
  • Multi-tenant competition
  • SLA constraints
  • Rising energy consumption
  • Escalating infrastructure costs

Containers offer lower overhead than virtual machines, but improper configuration, scaling, and scheduling can still lead to significant waste, redundancy, and energy inefficiency.

TPMCD introduces a sensitivity-driven hybrid container–VM scheduling and clustering framework that simultaneously optimizes:

• Cost
• Throughput
• Real Execution Time
• Number of Active Nodes

The method integrates clustering, intelligent threshold detection, scoring mechanisms, and hybrid configuration of containers and virtual machines.


🚨 The Core Problem

Cloud scheduling is inherently:

  • NP-Complete
  • Multi-objective
  • Dynamic under fluctuating workloads
  • SLA-constrained
  • Resource-competitive

Challenges include:

• Overloaded hosts
• Underutilized containers
• Excess VM provisioning
• Load imbalance
• Data redundancy
• Increased carbon footprint

Traditional methods treat:

  • Containers and VMs independently
  • Size-based classification only
  • Static thresholding
  • Single-objective optimization

TPMCD breaks that limitation.


💡 TPMCD Innovation

TPMCD (ThroughPut and Cost Optimizing Method for Clustering Tasks and Hybrid Containers in Cloud Data Centers) introduces:

1️⃣ Sensitivity Rate-Based Task Classification

Tasks are classified into low, medium, and high-load categories based on dynamic sensitivity metrics.

2️⃣ Hybrid Container–VM Configuration

Simultaneous optimization of:

  • Container allocation
  • Virtual machine assignment
  • Resource scoring
  • Communication efficiency

3️⃣ Intelligent Threshold Detector

Dynamic threshold adjustment prevents:

  • Overloading
  • Underloading
  • Critical imbalance points

4️⃣ Scoring-Based Resource Allocation

Each task is assigned based on:

  • Weight
  • Size
  • Sensitivity
  • Execution priority
  • Confidence interval estimation

5️⃣ Re-Clustering Mechanism

Reduces redundancy and improves real execution time.


🔬 Mathematical & Computational Model

TPMCD integrates:

  • Linear programming models
  • Sensitivity-based VM range mapping
  • Underload / overload detection formulas
  • Resource free-capacity estimation
  • Multi-objective cost functions

Objective Function simultaneously balances:

Minimize:

  • Total Cost
  • Real Execution Time
  • Node Utilization Overhead

Maximize:

  • Throughput
  • SLA Compliance

Simulation Environment:

  • Python 3.6
  • CloudSim 6.0

📊 Experimental Results

Compared to baseline scheduling methods:

• 7% average cost reduction
• 4% throughput improvement
• 9.5% real execution time reduction
• 3% fewer active nodes compared to KC method

Performance advantage increases under:

  • Dynamic workloads
  • High task concurrency
  • Multi-tenant cloud scenarios

🌱 Sustainability & Green Impact

Data centers significantly contribute to global energy consumption and greenhouse emissions.

TPMCD reduces:

  • Redundant containers
  • Excess VM rental
  • Resource wastage
  • Energy consumption

Hybrid container–VM regulation provides:

  • Better elasticity
  • Controlled autoscaling
  • Reduced idle overhead

🧠 Real-World Applications

  • Kubernetes scheduling optimization
  • Container orchestration systems
  • Hybrid cloud infrastructures
  • Enterprise SLA-based deployments
  • Scientific workflow processing
  • DevOps batch scheduling
  • Green cloud migration strategies

⚙ Technical Architecture Overview

Pipeline:

  1. Task classification via sensitivity rate
  2. Container assignment
  3. VM mapping within sensitivity ranges
  4. Threshold-based load adjustment
  5. Re-clustering for optimization
  6. Confidence interval validation

📈 Why This Research Matters in 2025

Cloud-native computing is container-driven.

Without intelligent scheduling:

  • Cloud waste increases
  • Carbon cost escalates
  • SLA violations grow
  • Infrastructure ROI decreases

TPMCD offers a scalable, hybrid, mathematically grounded framework for sustainable and cost-efficient cloud operation.


📄 Publication Details

Title: TPMCD: A method to optimizing cost and throughput for clustering tasks and hybrid containers in the cloud data center

Author: A. Ghorbannia Delavar

Journal: Journal of Network and Computer Applications

DOI: 10.1016/j.jnca.2025.104132

Year: 2025


📚 Citation (APA)

Ghorbannia Delavar, A. (2025). TPMCD: A method to optimizing cost and throughput for clustering tasks and hybrid containers in the cloud data center. Journal of Network and Computer Applications. https://doi.org/10.1016/j.jnca.2025.104132


📚 Citation (BibTeX)

@article{TPMCD2025, title={TPMCD: A method to optimizing cost and throughput for clustering tasks and hybrid containers in the cloud data center}, author={Ghorbannia Delavar, Arash}, journal={Journal of Network and Computer Applications}, year={2025}, doi={10.1016/j.jnca.2025.104132} }


🏫 Affiliation

Department of Computer Science
Payame Noor University, Tehran, Iran


⚖ License

This repository is created for academic indexing, discoverability, and research visibility purposes.
All publication rights belong to the journal publisher.


⭐ Digital Indexing Initiative (2022–2025)

This repository is part of a structured academic indexing program to improve global discoverability of high-impact research in:

• Cloud Computing
• Hybrid Scheduling
• Sustainable Datacenters
• Container Optimization
• Multi-objective Resource Allocation