Compile DAG

After a user has requested resources, resources may be paired with federated recipes or constructed in two new user-defined directed acyclic graphs (DAGs). These graph structures represent a flowchart that describes how data moves from logical component to logical component until the work is completed and data is stored in some destination. Great examples of this are NodeRed, GNU Radio, and Apache Airflow. The DAG data structure may be thought of as a symbolic representation of the work to be done. However, for actual work to be performed, this logical structure must be instantiated against physical components such as computational nodes and instrumentation. To compile a DAG, the symbolic representation is sent to an interpreter to build a job.

Preconditions

• the user is logged into the system

• the user has performed a resource query

• optionally the user has queried an appropriate federated recipe and potentially adapted it for their particular use case

• optionally, the user has constructed a new DAG from scratch

Post-conditions

• The DAG is passed to the interpreter to change the representation from a symbolic version to a job which has the effect of allocating the resources such that for some scheduled allotment of time all resources have been mutexed so that only the current user is capable of accessing them without a manual override.

• A smart contract has been compiled corresponding to the set of resources and the project has been billed for their use.