Conference paper Open Access

Multi-objective Container Deployment on Heterogeneous Clusters

Hu, Yang; de Laat, Cees; Zhao, Zhiming


JSON-LD (schema.org) Export

{
  "description": "<p>Operating system (OS) containers are becoming</p>\n\n<p>increasingly popular in cloud computing for improving productivity</p>\n\n<p>and code portability. However, existing deployment</p>\n\n<p>scheduling solutions mainly treat each container deployment</p>\n\n<p>as an independent request, and focus on the single aspect of</p>\n\n<p>resource utilization or load balancing, or work on homogeneous</p>\n\n<p>clusters. In this paper, we propose a new container deployment</p>\n\n<p>algorithm to satisfy multiple objectives on heterogeneous clusters.</p>\n\n<p>We analyze the deployment requirements of container-based</p>\n\n<p>infrastructure and formulate the deployment problem as a vector</p>\n\n<p>bin packing problem with heterogeneous bins. We focus on</p>\n\n<p>three objectives: multi-resource guarantee, load balancing, and</p>\n\n<p>dependency awareness. The goal of the proposed algorithm is</p>\n\n<p>to improve the tradeoff between load balancing and dependency</p>\n\n<p>awareness with multi-resource guarantees. Based on the algorithm,</p>\n\n<p>we implement a prototype scheduler to deploy containers</p>\n\n<p>on heterogeneous clusters. We evaluate our scheduler over a</p>\n\n<p>wide range of workload scenarios by simulation, which shows</p>\n\n<p>that our scheduler significantly outperforms existing schedulers</p>\n\n<p>of the container orchestration platforms.</p>", 
  "license": "https://creativecommons.org/licenses/by/4.0/legalcode", 
  "creator": [
    {
      "affiliation": "University of Amsterdam", 
      "@type": "Person", 
      "name": "Hu, Yang"
    }, 
    {
      "affiliation": "University of Amsterdam", 
      "@type": "Person", 
      "name": "de Laat, Cees"
    }, 
    {
      "affiliation": "University of Amsterdam", 
      "@id": "https://orcid.org/0000-0002-6717-9418", 
      "@type": "Person", 
      "name": "Zhao, Zhiming"
    }
  ], 
  "headline": "Multi-objective Container Deployment on Heterogeneous Clusters", 
  "image": "https://zenodo.org/static/img/logos/zenodo-gradient-round.svg", 
  "datePublished": "2019-10-01", 
  "url": "https://zenodo.org/record/3466785", 
  "version": "Camera ready", 
  "@type": "ScholarlyArticle", 
  "keywords": [
    "Container", 
    "Deployment", 
    "Multi-objective", 
    "Heterogeneous"
  ], 
  "@context": "https://schema.org/", 
  "identifier": "https://doi.org/10.1109/CCGRID.2019.00076", 
  "@id": "https://doi.org/10.1109/CCGRID.2019.00076", 
  "workFeatured": {
    "url": "https://www.ccgrid2019.org/", 
    "alternateName": "CCGrid19", 
    "location": "Cyprus", 
    "@type": "Event", 
    "name": "International workshop on Network aware big data computing, in the proceedings of IEEE CCGrid 2019"
  }, 
  "name": "Multi-objective Container Deployment on Heterogeneous Clusters"
}
127
57
views
downloads
Views 127
Downloads 57
Data volume 13.0 MB
Unique views 122
Unique downloads 51

Share

Cite as