wiki:CosmosOverview/Concepts

Version 7 (modified by msherman, 4 years ago) ( diff )

Concepts

This is a brief overview of the main concepts and terms needed to follow Cosmos Overview/Workflow

Resources

Everything that Experimenters can interact with is referred to as a Resource

To run an experiment, you will configure the resources that you need.

This includes:

  • Compute
    • capable of having disk images loaded and saved
    • users can log into them with ssh after provisioning
  • Software Defined Radios
    • FPGA images loaded with UHD
    • Radio paramaters controlled with UHD or other API
  • RF Frontends
    • api to control switches, filters, and gain
  • Network switches
    • connectivity defined with API
  • Optical Devices
    • space switch, roadms, and other devices controlled with optical config service

Note, some resources will fall into more than one of the above categories, such as a compute machine with integrated SDR, like the USRP 2974

Domains

  • Resources are grouped under Domains. Each domain won't interfere with any others, and they are reserved completely separately.
  • The hardware in a domain may span several sites, or may be contained within a single site.

Sites

  • Physical locations where hardware is installed.
  • A site consists of one or more nodes.

Nodes

  • Nodes refer to a set of radio, compute, and other resources meant to be used as a unit.
  • The resources in a node will share infrastructure such as power, networking, or RF frontends.
  • Therefore, the use of resources within a node must be coordinated to avoid interference.
  • It is possible to use a subset of the resources within a node

Reservations

  • Reservations allow users to obtain exclusive use of resources for a define time period.
  • Currently, to ensure repeatable results, resources are reserved at the domain level.
  • TODO Once the framework is in place, it will be possible to reserve a combination of nodes + frequency allocation, but for now the time slot guarantees exclusive use.

Images

  • Compute resources have their state defined by what disk image is loaded
  • Predefined base images are available, and described at User Guide/Image Versions
  • Tutorials will either use a base image, or a specific customized image.
  • The normal workflow is to load a base image, make the changes you need, and then save your own customized image to preserve your work.
  • Details on how to work with images are at OMF Quick Start

Tutorials

  • Tutorials are used as both instructional tools, as well as starting points for developing your own experiments.
  • For example, following the basic usage tutorial, you will be able to send and receive RF data with a SDR.
Note: See TracWiki for help on using the wiki.