Failure Prediction with Hierarchical Approach in Private Cloud

Yaru Bao, Feilong Tang, Lijun Cao

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Cloud computing is widely adopted in real-world data centers. Most companies choose to build a private cloud service with the consideration of privacy. In these circumstances, they provide the service through Infrastructure as a Service (IaaS). However, with the scale of the data center, the possibility of cloud failure is increasing and become urgent in cloud computing. Current methods mainly use the proactive approach that monitors the failure and process it afterward. These methods are inefficient, and may always cause the service to break down. In this paper, we propose a new approach HFP (Hierarchical Failure Prediction) that can effectively monitor and predict the failure in advance. We firstly design and implement a new monitor structure that can effectively collect data. Then, a failure prediction method is proposed to predict the failure in advance. We implement this system with OpenStack, the synthetic result on the collected dataset shows that our method can achieve 98.3% accuracy in the prediction of cloud failure.

Original languageEnglish
Title of host publicationGreen, Pervasive, and Cloud Computing - 15th International Conference, GPC 2020, Proceedings
EditorsZhiwen Yu, Christian Becker, Guoliang Xing
PublisherSpringer Science and Business Media Deutschland GmbH
Pages469-480
Number of pages12
ISBN (Print)9783030642426
DOIs
StatePublished - 2020
Externally publishedYes
Event15th International Conference on Green, Pervasive, and Cloud Computing, GPC 2020 - Xi'an, China
Duration: 13 Nov 202015 Nov 2020

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12398 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference15th International Conference on Green, Pervasive, and Cloud Computing, GPC 2020
Country/TerritoryChina
CityXi'an
Period13/11/2015/11/20

Keywords

  • Cloud computing
  • Failure prediction
  • Machine learning

Fingerprint

Dive into the research topics of 'Failure Prediction with Hierarchical Approach in Private Cloud'. Together they form a unique fingerprint.

Cite this