Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.
Comment: Adjusted the format, used "Heading 2"

...

  • Seems to focus on streamed execution of legacy software.
  • Has overlap with DAP.  Only client software is available, server software not released yet. Might be usable for extractors that require a special OS/software environment.  Software only available for Linux. Software may not be mature enough.  June 2013 presentation listed "thin-client execution in OpenStack" as a future work item.

 

What is it

...

A collaborative (originally between Carnegie Mellon University (CMU) and IBM Research) project seeking to establish a robust ecosystem for
long-term preservation of software, games and other executable content.  Born at CMU, Olive addresses the current gap in preservation technology by providing a curated environment for the preservation and distribution of executable content.
-- From olivearchive.org

Executable content examples: simulation models, tutoring systems, expert systems, data visualization tools, company's product information page that are dynamically generated that customizes the content and appearance at run time.

Goals

- Enable libraries to achieve their mission of preserving scholarly records
- Provide a platform for preserving, searching, extending, and distributing executable content
- Establish a workflow for this preservation
- Determine the obstacles to preservation and distribution
- Retain executable content in its original state so that end users may interact with it freelyHow It Works
- Executable content is preserved in a full VM to ensure long-term compatibility
- Relevant VM archives can be found using content and metadata searches
- VMs are streamed using the ISR technology.
  Like Internet videos, users may begin interacti  ng before a full VM is downloaded. Streaming technology offers immediate VM access by sending only the data required for execution.
- Users may save VM changes to their own computers and apply them to subsequent launches.
Use cases:
- Academic
  Executable journal articles, complex datasets, research tools and
  content, databases and navigation tools, software and its
  dependencies.
- Public Sector
  Modify legacy software on old hardware for:
  Government databases and datasets, security technology, intelligence
  software and systems, special-use hardware preservation.
- Private Industry
  Help maintain legacy hardware and software for clients in sensitive
  hardware configurations, such as:
  Game preservation for future development or analysis, software
  execution on deprecated systems, system testi  ng and maintenance for
  legacy infrastructure, stable and controlled environment for special
  hardware and software.
Software
- Client: VMNetX.
  Runs on Linux, using the open-source KVM VM monitor.  Can execute
  VMs directly from any web server; no special server software is
  required. On GitHub. License: GPL v2.
https://github.com/cmusatyalab/vmnetx
- Server: not released, plan to release under GPL v2.
Client releases: VMNetX
0.4.4 - bugfix release, 2014-05-31.
0.4.3 2014-03-05.
0.4.2 2013-12-20.
0.4.1 2013-11-07.
Funding:

 

Team

- Prof. Mahadev Satyanarayanan (Satya), in CS, CMU

  Architect of Andrew, Coda files systems

- Gloriana St. Clair

  Dean Emerita of University Libraries, CMU

- Erika Linke

  Associate Dean of University Libraries, CMU

- Jerome McDonough

  Associate Prof, UIUC

- Anita de Waard

  VP of Research Data Collaborations, Reed Elsevier

- Daniel Ryan

  Curator for Executable Content, CMU

...

 

 

Funding

- IMLS National Leadership Grant (Institute of Museum and Library
  Services, a government agency)

  Oct 2012, $497k over 2 years.

  Goal: To understand what types of content can be ingested into Olive
  and to determine how executable content archive can fit into
  existing trusted repository standards such as OAIS, OCLC, CRL, JISC.

- Sloan Foundation Grant.

  Jan 2013. $400k over 2 years.

  Goal: To develop the technical framework for Olive, to plan for an
  effective organizational structure to sustain the archive and
  provide access to executable content.

 

Test cases

...

- Great American History Machine

  Built by Prof. David Miller at CMU in the 1980’s, provides GIS for early census data

- ChemCollective

  Award winning educational software built by Prof. David Yaron

- Mystery House

  Apple II game for which experimental metadata was generated under IMLS-funded Preserving Virtual Worlds 2

- Doom

  Popular PC game with heavy graphics released in 1993

- Information Systems

  Special issue of the Elsevier journal which will include executable contentCMU team:
- Prof. Mahadev Satyanarayanan (Satya), in CS
  Architect of Andrew, Coda files systems
- Gloriana St. Clair
  Dean Emerita of University Libraries, CMU
- Erika Linke
  Associate Dean of University Libraries, CMU
- Jerome McDonough
  Associate Prof, UIUC
- Anita de Waard
  VP of Research Data Collaborations, Reed Elsevier
- Daniel Ryan
  Curator for Executable Content, CMU
...
Three (3) ideas:


How It Works

- Executable content is preserved in a full VM to ensure long-term compatibility

- Relevant VM archives can be found using content and metadata searches

- VMs are streamed using the ISR technology.
  Like Internet videos, users may begin interacti  ng before a full VM is downloaded. Streaming technology offers immediate VM access by sending only the data required for execution.

- Users may save VM changes to their own computers and apply them to subsequent launches.

Software

- Client: VMNetX.
  Runs on Linux, using the open-source KVM VM monitor.  Can execute VMs directly from any web server; no special server software is required. On GitHub. License: GPL v2.  https://github.com/cmusatyalab/vmnetx
- Server: not released, plan to release under GPL v2.

Client releases: VMNetX
0.4.4 - bugfix release, 2014-05-31.
0.4.3 2014-03-05.
0.4.2 2013-12-20.
0.4.1 2013-11-07.

Three main ideas

- ISR (Internet Suspend/Resume).  A technology to stream VMs over
  Internet. -- From CMU.
  Disk is demand-paged over the network.
- A technique for indexing and searching VM image contents.  Searching
  relies on signature matching, instead of user-specified metadata.
- A technique to incrementally composing a VM image.