Storage Needs

From Dryad wiki
Jump to: navigation, search

Frequently, Dryad needs to estimate storage requirements for the future. This page collects the most recent calculations.

Overview

Dryad needs to store many types of information (temporary working files, this wiki, metadata database, source code, etc.). However, the largest factor by far is the storage needed for the actual data files submitted to the repository. Therefore, these calculations will only deal with the data files.

Current use

The assetstore is mirrored on each Dryad machine. For machines that run multiple instances of Dryad, multiple copies of the assetstore may be needed.

Historical size of the Dryad production data (/dryad-data volume) can be found in the dataStorage.csv on GitHub.

Bandwidth:

  • 2014-4-3, due to lots of CLOCKSS harvesting activity, we had 151745 downloads in the previous month. This is roughly 10 times the current content.

Projected use

  • 2014-3-6 -- the number of data packages is roughly doubling each year, and the average size of a data package is also doubling each year.
  • 2012-7-15 -- we estimated that Dryad's needs would double yearly. This gives us a 5 year projection of 4TB.
  • In 2010, For the DryadUK proposal, we estimated that Dryad would need up to 1TB of space over 3 years (up to fall 2013). This will allow us to scale up to roughly 10 times the number of journals we have now, assuming the deposit rates are relatively consistent.