Return to search

Slimming virtual machines based on filesystem profile data

Virtual machines (VMs) are useful mechanisms for better resource utilization, support for special software configurations, and the movement of packaged software across systems. Exploiting VM advantages in a production setting, however, often requires computer systems with the smallest possible disk-size footprint. Administrators and programmers who create VMs, however, may need a robust set of tools for development. This introduces an important conflict: Minimalism demands that packaged software be as small as possible, while completeness demands that nothing required is missing. We present a system called Lilliputia, which combines resource usage monitoring (through a Linux FUSE filesystem we created called StatFS), with a filtered cloning system, which copies an existing physical or virtual machine into a smaller clone. Finally, we show how Lilliputia can reduce the size of the Trellis Network-Attached-Storage (NAS) Bridge Appliance and the Chemical Shift to 3D Structure protein structure predictor to 10-30% of their original size.

Identiferoai:union.ndltd.org:LACETR/oai:collectionscanada.gc.ca:AEU.10048/1489
Date11 1900
CreatorsNickurak, Jeremy
ContributorsLu, Paul (Computing Science), Galin, Warren (Biological Sciences), Nikolaidis, Ioanis (Computing Science)
Source SetsLibrary and Archives Canada ETDs Repository / Centre d'archives des thèses électroniques de Bibliothèque et Archives Canada
LanguageEnglish
Detected LanguageEnglish
TypeThesis
Format519922 bytes, application/pdf

Page generated in 0.0021 seconds