There are a number of different types of wave servers within the NCSS.

The wave servers that most applications (jiggle, tmts, waveform archivers) will use are proxy wave servers (PWS). PWS does not itself serve any waveforms. Instead, it provides a protocol translation service between the client and the back-end wave servers. Applications speaking the CISN waveclient protocol send requests (SNCL, start and end times) to PWS. PWS looks up the requested SNCL in a database table to determine which type of back-end server can service the request. The PWS queries the one or more back-end servers that are configured to handle that request for the PWS. The supported types of back-end servers that PWS can handle are Earthworm wave_serverV (including Winston wave servers), Nanometrics NaqsServer, and CISN Simple Wave Server (SWS). PWS uses the appropriate protocol to request waveforms from the back-end servers. In the event that the waveform returned from the first back-end server does not complete the request, PWS will set the returned waveform aside and query a second back-end server if available. If the returned waveform more completely satisfies the request than did the first waveform, then the second will replace the first. This process continues until either the waveform completely satisfies the request or there are no more back-end servers that serve the requested SNCL. Finally the waveform is returned to the client. Note that PWS makes no attempt to merge the partial waveforms from two back-end servers into a single waveform: there is no way to do that in general.

There are several different proxy wave servers on the primary NCSS computers, each one configured for a specific purpose. The following tables outlines the various PWS servers and their uses.

Each of these Proxy Wave Servers is accessible from the Internet, the CISN net, and the NCSS net.