How does FAH work?

From FaHWiki

Jump to: navigation, search

General

The software involved in the Folding@home project consists of many collaborating parts: several kinds of servers and a variety of clients. Folding@home uses a client-server model where the clients receive a work unit (WU) from the servers, process it and return the processed results back to the servers where the results are analyzed.


The general sequence of events is:

  1. a client sends an assignment request to one of the Assignment Servers
  2. the Assignment Server assigns the client to a Work Server
  3. the client downloads the assigned WU from the Work Server and starts processing it


When the client is done processing the WU, the sequence of event is:

  1. the client attempts to upload the WU to the Work Server from which it downloaded the WU,
    • if the Work Server is unable to accept the WU, the client attempts to send the WU to the Collection Server
      • if the Collection Server is also unable to accept the WU, the client will keep the WU in its queue and will attempt to upload it at periodic intervals
  2. the WU is credited and the points are added to the user configured on the client who sent it


After a WU is uploaded by a client and depending on the results, further WUs are generated based on the uploaded WU and made available for assignment to a client.

Personal tools