CHAVE Project

Consolidation and High Availability on Virtualized Environments

Cloud Computing Virtualization Energy Efficiency High Availability Virtual Machine Management

Project Overview

CHAVE is a research project that proposes an innovative approach to virtual machine management that balances energy efficiency with high availability in cloud computing environments.

The project addresses a fundamental challenge in cloud infrastructure: how to reduce energy consumption through virtual machine consolidation while maintaining sufficient service availability. By developing new algorithms and metrics, CHAVE provides a framework for optimizing resource allocation in virtualized environments.

Key Innovations

  • Energy-Aware VM Placement Algorithm

    Novel placement strategies that consider both power consumption and performance metrics

  • Availability Risk Assessment

    Mathematical models to quantify the risk of service unavailability during VM consolidation

  • Dynamic Workload Analysis

    Real-time monitoring and prediction of VM resource demands to optimize placement decisions

  • Multi-Objective Optimization

    Balancing conflicting goals of energy efficiency, performance, and availability

  • Migration Cost Modeling

    Assessment of the overhead and impact of VM migrations on system performance

Technical Architecture

The CHAVE framework consists of several interconnected components:

Core Components
  • Resource Monitor - Collects and analyzes VM and host metrics
  • Consolidation Engine - Implements VM placement algorithms
  • Risk Analyzer - Evaluates availability risks
  • Decision Manager - Orchestrates VM migrations
  • Performance Validator - Ensures SLA compliance
Integration Points
  • OpenStack and VMware adapters
  • Power measurement interfaces
  • Workload prediction engine
  • Historical data analytics
  • Event notification system

Research Outcomes

The project has produced significant contributions to the field of cloud resource management:

CHAVE: Consolidation with High Availability on Virtualized Environments

Master's Dissertation, UDESC, 2018

Download Dissertation
Uma Proposta de Orquestração de Nuvem Computacional Baseada em Consolidação, Elasticidade e Disponibilidade

ERAD-RS (2017)

Download Paper
MeHarCEn: A Method of Harmonizing Energy Consumption in Data Centers

RITA (2017)

Download Paper

Experimental Results

Tests conducted using OpenStack cloud environments demonstrated that the CHAVE approach can:

25%

Energy reduction compared to traditional VM allocation approaches

99.95%

Service availability maintained while optimizing resource usage

Project Snapshot
  • Status: Completed
  • Duration: 2016 - 2018
  • Role: Principal Researcher
  • Supervisor: Prof. Dr. Charles Miers
  • Source Code:
    GitHub Repository
Technologies Used
Python OpenStack VMware Docker Java Machine Learning Predictive Analytics NoSQL Network Simulation Statistical Analysis
Related Projects

Related Publications

Uma ferramenta para análise do impacto da migração de máquinas virtuais

REC (2017)

Analysis of the impact of virtual machine migration on network and system performance, considering different hypervisors and VM configurations.

Download Paper
Consolidação de máquinas virtuais com alta disponibilidade: uma revisão bibliográfica sistemática

MEP (2017)

Systematic literature review on virtual machine consolidation strategies that maintain high availability in cloud environments.

Download Paper
Análise da consolidação de máquinas virtuais nos equipamentos de rede

SDA (2016)

Analysis of the impact of virtual machine consolidation on network equipment and traffic patterns in data center networks.

Download Paper
Problema Bin Packing com Abordagem Heurística First Fit Decreasing

PAA (2016)

Implementation and analysis of the First Fit Decreasing heuristic for the Bin Packing problem, applied to virtual machine placement.

Download Paper

Get in touch

Interested in learning more about the CHAVE project?

Please provide your name.
Please provide a valid email.
Please provide a message.