Using accelerated push distribution

Enabling accelerated push greatly speeds up distribution tasks. The default push distribution task type builds target lists of up to 241 devices and processes these devices one target list at a time, waiting until jobs complete before moving on to devices in the next target list.

Accelerated push distribution makes this process asynchronous. As the core discovers and communicates with target devices, it tells them what to do and then moves on to the next targeted device without waiting for the job to complete. This discovery and communication process uses multiple processor cores and threads. Each device then processes the job on its own and sends job status to the core server when necessary.

Accelerated push scheduled tasks use a single target list internally, and by default accelerated push processes up to 64 targets concurrently from the list. You can adjust this if necessary based on your core server resources. Generally you won't need to change the default.

Accelerated push also works well with multicast. Accelerated push contacts target devices much more quickly than normal push jobs, so a multicast distribution can include even more devices at once. The device acting as the multicast representative in each subnet waits the configured amount of time before beginning the multicast and all other devices using that representative will receive the multicast once it starts. Any devices late to the multicast will still get what was left of the multicast when they joined and then they will backfill what they missed from their peers or the source.

Supported task types:

  • Policy-supported
  • Policy
  • Push

Supported download options:

  • Run from source (execute on share)
  • Download and execute
  • Pre-cache—just download the package or bundle without installing.

Supported platforms:

  • macOS
  • Windows