File mirroring scenario
Download as PPT, PDF0 likes205 views
This document discusses mirroring a collection of files from a central archive to a local community repository. It proposes building a machine at the central archive to test mirroring and load the data, then ship it to the community. The community could then build their own front-end using the mirrored data, either developing their own interface or using Dspace. Updates from the community could then be periodically mirrored back to the central archive. Inspiration is drawn from work done by Tim Hopkins at the University of Kent to develop robust mirroring technology using turn-key packages that allow any site to set up their own mirror service on cheap hardware.
![File Mirroring Scenario Doug Moncur [email_address]](https://image.slidesharecdn.com/filemirroringscenario-100707230711-phpapp02/85/File-mirroring-scenario-1-320.jpg)
File mirroring scenario
- 1. File Mirroring Scenario Doug Moncur [email_address]
- 2. What’s a mirror? Internet mirror is a way of faithfully copying an archive Robust copying techniques to ensure copy is a faithful copy Long history – designed to cope with poor communication lisks
- 3. Scenario AIATSIS holds a collection of material specific to a community Files have unique identifier Mirror collection to community ► Community gets data not front end ► Community needs to build own front end
- 4. Build own front end? Needs to either develop own interface Or use something like Dspace to build local archive using mirrored data ► Build machine at AIATSIS say, test mirroring, load data, ship to community
- 5. Benefits Data stays safe in central repository Community has local access Web based technology – can easily share data within knowledge centre Builds from public domain/open source software No licensing issues or costs
- 6. Extensions Community can add material to collection New material and data could be periodically mirrored back to central archive ► No need for local back up (Implications re communications and communication charges)
- 7. Inspiration Tim Hopkins, University of Kent in UK has been working on mirroring technology for years Responsible for development of original UK Higher Education Software archive ( www.mirrorservice.org)
- 8. Inspiration continued From an email … The second thing we're going to look at is turning the complete miorror kit into a turn-key package. Once that's done (and we're quite a way towards it) any site can set up it's own mirror service and this opens up loads of possibilities as far as co-operative working is concerned. For example, mirrors could take copies from the nearest (geographically) available source rather than the origin. The bottom line as far as your project is concerned is that you could set up distributed networks of mirrors and keep duplicate copies of the archive spread across an arbitrary number of sites. In addition, the syatem would run on cheap hardware. Can build simple robust solution on cheap hardware!