TU Darmstadt / ULB / TUbiblio

Long-Term Preservation of Data Analysis Software with Operating-System-Level Virtualization

Eichhorn, Helge and Trinkel, Thomas and Anderl, Reiner (2015):
Long-Term Preservation of Data Analysis Software with Operating-System-Level Virtualization.
In: PV Conference, Darmstadt, 3.-5. November 2015, [Conference or Workshop Item]

Abstract

A lot of today's computer-aided engineering tasks, such as data analysis, are computationally expensive, highly domain- or problem-specific, and of high complexity. Due to this fact commercial off-the-shelf software solutions do not satisfy the requirements of many organizations and custom software tools are developed. For the sake of knowledge management and traceability it should be possible to reproduce and modify such analyses with minimal effort in future engineering processes. It is not sufficient though to archive source code together with input data since every program generally relies on an extensive dependency tree. An emerging trend in the cloud computing industry is the move towards OS-level virtualization technologies, so-called containers. Within containers applications and their dependencies are packaged. We propose a workflow for archiving of custom software tools based on a hybrid virtualization strategy. A real world example is developed and implemented with the Docker container platform. Finally performance overhead and storage requirements are compared to VM-based solutions.

Item Type: Conference or Workshop Item
Erschienen: 2015
Creators: Eichhorn, Helge and Trinkel, Thomas and Anderl, Reiner
Title: Long-Term Preservation of Data Analysis Software with Operating-System-Level Virtualization
Language: English
Abstract:

A lot of today's computer-aided engineering tasks, such as data analysis, are computationally expensive, highly domain- or problem-specific, and of high complexity. Due to this fact commercial off-the-shelf software solutions do not satisfy the requirements of many organizations and custom software tools are developed. For the sake of knowledge management and traceability it should be possible to reproduce and modify such analyses with minimal effort in future engineering processes. It is not sufficient though to archive source code together with input data since every program generally relies on an extensive dependency tree. An emerging trend in the cloud computing industry is the move towards OS-level virtualization technologies, so-called containers. Within containers applications and their dependencies are packaged. We propose a workflow for archiving of custom software tools based on a hybrid virtualization strategy. A real world example is developed and implemented with the Docker container platform. Finally performance overhead and storage requirements are compared to VM-based solutions.

Uncontrolled Keywords: data analysis, container-based virtualization, docker, software archiving
Divisions: 16 Department of Mechanical Engineering > Department of Computer Integrated Design (DiK)
16 Department of Mechanical Engineering
Event Title: PV Conference
Event Location: Darmstadt
Event Dates: 3.-5. November 2015
Date Deposited: 19 Nov 2015 15:51
Export:
Suche nach Titel in: TUfind oder in Google
Send an inquiry Send an inquiry

Options (only for editors)

View Item View Item