ViSUS configuration file

From
Revision as of 20:08, 18 August 2017 by Admin (talk | contribs) (Access)
Jump to: navigation, search

Configuration File Location

The software's current strategy is to search the following locations in this order:

  1. location specified by --visus-config command line parameter
  2. current working directory
  3. visus user directory (note: itunes-accessible documents directory on ios)
  4. resources directory (e.g. osx/ios bundle, same as cwd on other platforms)
  5. hardcoded path provided at compile time (e.g. for mod_visus default)
  6. executable directory (very unusual deployment, but it doesn't hurt anything)

Configuration Options

A simple ViSUS configuration file would look like the following:

<visus>
    <!-- Local dataset --> 
    <dataset name="my_dataset" url="file:///Users/username/data/my_dataset.idx" permissions="public"/>

    <!-- Remote dataset -->
    <dataset name="2kbit1" url="http://atlantis.sci.utah.edu/mod_visus?dataset=2kbit1" permissions="public">

    <!-- Saved scene -->
    <scene url="file:///Users/username/data/my_scene.xml" permissions="public"/>
</visus>

Access

ViSUS can access the same data from both the network or a local cache on your disk. The cache allows faster (and offline) access to the data that already navigated using the viewer. Please notice that ViSUS only saves the timesteps and fields that you explored, not the entire dataset.

An important feature of the ViSUS framework is the caching. Data streamed from the server can be cached on disk for later faster access. To enable this feature for a specific dataset we can define a dataset in the configuration file as following:

  <dataset name="2kbit1" url="http://atlantis.sci.utah.edu/mod_visus?dataset=2kbit1" permissions="public">
     <access name="Multiplex" type="multiplex">
       <access name="cache"  type="disk" chmod="rw" url="$(VisusCacheDirectory)/2kbit1/visus.idx" />
       <access name="source" type="network" chmod="r" compression="zip" />
     </access> 
  </dataset> 

With these lines of code we created two layers of data access the first from the disk (e.g. the cache) and the other from the network. This will allow the application to search first on the disk for the requested data and in case of miss the request will be forwarded to the network. At the same time any data requested to the network layer will be cached on disk (at the specified URL).

Note: when frequent and fast access is required for specific dataset, enabling the caching is highly recommended.

Faster network access In order to speed up your network performance it is possible to use some special options in your access configurations:

  • nconnections
  • num_queries_per_request

This will enable parallel requests and gather multiple block requests in a single query.

Here is an example:

<access name="source" type="network" chmod="r" compression="zip" nconnections="4" num_queries_per_request="32"/>