A job network is a group of jobs that stand in defined relation to each other. This relation is composed of dependencies, which are expressed as logical conditions. In the simplest case, two jobs in a job network can be linked by the condition: If Job 1 finishes OK, start Job 2 (see Example of Job Linkage by Using Conditions).

A job network is uniquely defined by its owner and network name. Each network is given a start and deadline time which determine when the network is to be activated. If your installation includes multi-CPU support, you can also specify a default node name for the jobs in the network. This node name can be overridden at the job level (see Operating System Server Nodes in the Concepts and Facilities documentation).

A user can only access a defined job network if the user's ID is associated with the same owner as the network, unless the user has special authorization to access other networks (see Granting Definition: Authorizing Other Users or Owners to Access a Network).

A job network and a single job are the units of work that can be activated by Entire Operations. When a job network is activated, it is automatically given a run number that uniquely identifies this network activation. This feature allows several copies of the same job network to run simultaneously.

A job network can be a subnetwork of another job network.