Abstract
Virtualisation technology has become a very common trend in modern datacentres as Virtual Machine (VM) migration brings several benefits like improved performance, high manageability, resource consolidation and fault tolerance. Live Migration (LM) of VMs is used for transferring a working VM from one host to another host of a different physical machine without interfering with the existing VMs. However, little research has been done in considering the real time resource consumption and latency of live VM migration that reduces these benefits to much less than their potential. In this paper, we present an analysis of LM in our unique TransAtlantic high speed optical fibre network connecting Northern Ireland, Dublin and Halifax (Canada). We show that the total migration times as well as total network data transfer for post-copy LM are both dominated by specific VM memory patterns using loaded or unloaded VMs. We also found that the downtime for different VM memory patterns is not extremely varied and no severe effect is experienced over our long distance network.
Original language | English |
---|---|
Title of host publication | Proceedings of 2016 SAI Computing Conference, SAI 2016 |
Publisher | The Institute of Electrical and Electronics Engineers (IEEE) |
Pages | 1267-1272 |
Number of pages | 6 |
ISBN (Print) | 9781467384605 |
DOIs | |
Publication status | Published - 29 Aug 2016 |
Event | 2016 SAI Computing Conference, SAI 2016 - London, United Kingdom Duration: 13 Jul 2016 → 15 Jul 2016 |
Conference
Conference | 2016 SAI Computing Conference, SAI 2016 |
---|---|
Country/Territory | United Kingdom |
City | London |
Period | 13/07/16 → 15/07/16 |
Keywords
- Cloud
- Live Migration
- Openstack
- Optical Network
- Virtual Machine
Profiles
-
Gerard Parr
- School of Computing Sciences - Professor of Computing Sciences
- Cyber Security Privacy and Trust Laboratory - Member
- Data Science and AI - Member
- Smart Emerging Technologies - Member
Person: Research Group Member, Academic, Teaching & Research