Chapter Contents

Previous

Next
SAS/CONNECT User's Guide

Distributed Agent Processing

Using compute services and distributed messaging, agents can be used to distribute work across nodes on a network that is most efficient for a client/server application. This distributed service was designed for client/server applications which benefit from unattended, asynchronous execution.

Agents transparently use SAS/CONNECT to sign on to a designated location using a SAS spawner and a designated user identity, and then execute the agent's source code. As agents are processed, the log and output windows are spooled until the agent has finished executing. The agent then signs off the designated node. Upon completion, a monitoring client may retrieve the spooled execution log and output, merge it with the local output or purge it.

In addition, agents may be designed to use the indirect-messaging facility to communicate information to client applications by using message queues. This allows agents to be constructed to generate alerts and/or dynamic reports as they execute. These conditions and information are then manifested as messages and written to queues, which are monitored by interested clients.

The initiating client SAS session is not required to remain active while the agent executes (unlike a conventional local or remote SAS/CONNECT session configuration).


Chapter Contents

Previous

Next

Top of Page

Copyright 1999 by SAS Institute Inc., Cary, NC, USA. All rights reserved.