deploy: ea9c4ae

edex/distributed-computing/index.html

@@ -1075,8 +1075,8 @@ <h1 id="distributed-edex">Distributed EDEX<a class="headerlink" href="#distribut
 <p>AWIPS makes use of service-oriented architecture to request, process, and serve real-time meteorological data. While originally developed for use on internal NWS forecast office networks, where operational installations of AWIPS could consist of a dozen servers or more, the early Unidata releases were stripped of operations-specific configurations and plugins, and released as a standalone server. This worked, since (at the time) a single EDEX instance with an attached SSD could handle most of NOAAport. However, with GOES-R(16) coming online in 2017, and more gridded forecast models being created at finer temporal and spatial resolutions, there is now a need to distribute the data decoding across multiple machines to handle this firehose of data.</p>
 <hr />
 <h2 id="unidatas-current-edex-server">Unidata's Current EDEX Server<a class="headerlink" href="#unidatas-current-edex-server" title="Permanent link"></a></h2>
-<p>Currently, with our specific EDEX server we use a Database/Request instance that also decodes and ingests a good portion of the data.   It handles all data requests from CAVE users, as well as the majority of the decoding and ingesting for data feeds coming down on the LDM.  The <strong>radar</strong> data has been specifically exluded (from the decoding and ingest) and it has its own <a href="#ingestdecode-server"><strong>Ingest/Decode Server</strong></a> which is explained in more detail below.</p>
-<p>For our EDEX we have designated an instance of the ingest/decoding server to be dedicated to handling the radar data.  Our <em>Radar-EDEX</em> recieves and decodes all radar down from the LDM and then stores it back on our main <a href="#databaserequest-server"><strong>Database/Request EDEX</strong></a> in the form of HDF5 data files and PostgreSQL metadata.</p>
+<p>Currently, we use a distributed architecture comprised of three machines: 1 <strong>main EDEX</strong> machine and 2 <strong>ancillary EDEX</strong> machines.  The main EDEX machine decodes and processes the majority of the data, while serving and storing all of the data.  Our two ancillary machines -- one for radar data and one for satellite data -- each decode and process a subset of the data and send it back to the main EDEX for storage and requesting.</p>
+<p>The main EDEX is an instance of our <a href="#databaserequest-server">Database and Request Server</a> and more information on our <a href="#ancillary-edex-server-ingestdecode-edex-server">ancillary EDEX machines</a> is below as well.</p>
 <hr />
 <h2 id="example-installation">Example Installation<a class="headerlink" href="#example-installation" title="Permanent link"></a></h2>
 <p>This walkthrough will install different EDEX components on two machines in the XSEDE Jetstream Cloud, the first is used to <strong>store and serve</strong> while the second is used to <strong>ingest and decode</strong> data.</p>