hybrid-cloud

Hybrid Cloud

An enterprise runs sensitive workloads on-premise and burst workloads in the public cloud. The hybrid pipeline needs to classify each task by data sensitivity, route sensitive tasks to on-premise workers and burst tasks to cloud workers, and aggregate results regardless of where they executed.

Pipeline

[hyb_classify_data]
     |
     v
     <SWITCH>
       |-- onprem -> [hyb_process_onprem]
       |-- cloud -> [hyb_process_cloud]

Workflow inputs: dataId, dataType, payload

Workers

HybClassifyDataWorker (task: hyb_classify_data)

Classifies incoming data by sensitivity. Checks dataType against a hardcoded sensitive-type list containing "pii", "phi", and "financial". If matched, sets classification = "confidential" and target = "onprem". Otherwise sets classification = "public" and target = "cloud". The target output drives the downstream SWITCH task.

• Reads dataType. Writes classification, target

HybProcessOnpremWorker (task: hyb_process_onprem)

Processes confidential data on the organization's own infrastructure. Reports processedAt = "datacenter-1" and encrypted = true, keeping sensitive data within the physical security boundary.

• Writes processedAt, encrypted

HybProcessCloudWorker (task: hyb_process_cloud)

Processes non-sensitive data in the public cloud for elastic scaling. Reports processedAt = "aws-us-east-1" and scaledInstances = 4, indicating horizontal scaling was applied.

• Writes processedAt, scaledInstances

---

12 tests | Workflow: hybrid_cloud_demo | Timeout: 60s

See RUNNING.md for setup and usage.