At Elastacloud, we have an expectation of building leading edge data stacks for large enterprise customers. We choose the best technologies that fit our ethos of elasticity, focus, agility and economy in the cloud.
A key emergent trend in our customer is the notion of a wide data stack.
Data acquisition is the role of the data stack and the data stack much pull data into a pipeline, not respond as a push. Where eventing approaches are taken, the event source invokes an endpoint that the data stack owns.
Data is free to migrate between media in the stack, whilst being controlled by the overall data governance process.
Compute is applied over Data with great fluidity and flexibility. To support this, the wider stack supports execution in a family of compute environment and languages; from Spark to USQL to R.
To achieve this at Elastacloud we have a runbook;
Everyone has to be a polymath; from ML to BI
Data has to be governed
Access to data has to be controlled but free