Alluxio is available via Docker. You can create a cluster of Alluxio within a Kubernetes cluster. Given that we do have these containers, you can either use a daemon set or a replica set within a Kubernetes cluster to create an alluxio cluster itself and have it co-located within your other nodes that may be running Presto or Spark. This is a way to create a data layer within kubernetes that’s memory based and you can still maintain really tight data locality for your Spark and Presto workloads.
Just like Kuberentes orchestrates compute and containers, Alluxio orchestrates data, brings it all together, and scales it across all the workers for compute – we think it’s a good analogy.