This document captures the things that would be needed to make requirements for making AI actions a first-class citizen in Openwhisk. The main ideas captured here bellow are inspired by from the work done to run in running deep learning inferences on images and videos.
...
The volume my-action-volume
is can be mounted inside each container on: /mnt/ow-tmp/my-action-volume.
TBD When the The volume should be removed once the prewarmed actions are destroyed.
Open questions:
Should the developers specify a size for the volume, or assume a default for each volume ?
Cluster cache
Provide developers with a caching solution in the cluster. Developers would still pass larger assets by reference between actions, they would still write code to upload or download an asset, but use a cache provided by OpenWhisk, inside the cluster. The cache can be exposed through the Openwhisk SDK through 2 methods: write(asset_name, asset_value)
, read(asset_name)
.
...