Oracle ⇄ ADLS Bridge

Modernize your Oracle archive strategy. This solution extracts data from
ANY TCP-reachable Oracle database — on-premises (via ExpressRoute/VPN),
OCI, AWS RDS, or Azure VM — and writes it to Azure Data Lake Storage
Gen2 as Snappy-compressed Parquet. Two-way: restore back to Oracle when
you need it, or query archived data directly via Synapse Serverless SQL
or Microsoft Fabric Lakehouse without recreating the database.

KEY CAPABILITIES

- Source-agnostic Oracle reads via JDBC over any TCP-reachable connection.
Three-phase parallel extract (PK / ROWID / LOB) with automatic table
classification and per-table SCN consistency.
- Captured DDL alongside data. Restore reapplies the schema with
portability cleanups (strips tablespace/storage clauses) and validates
row counts post-restore.
- Restore-to-query. Every extract emits Synapse + Fabric catalog SQL —
customers run one script and get queryable views over the Parquet
WITHOUT restoring to Oracle.
- HIPAA-ready. Optional private endpoints on storage and Key Vault, RBAC-
only access, and an optional 7-year time-based WORM immutability policy
for compliance archive.
- GDPR data-residency by design. Single-region deployment; archived data
never leaves the Azure region you choose at deploy time. Pair with
Azure Policy `allowedLocations` if you need tenant-wide enforcement.
- Tenant isolation. Everything runs inside the customer's Azure
subscription — no KinetiStack™-hosted service is in the data path,
ever.
- Lifecycle tiering. Standard profile transitions blobs Hot → Cool @ 30d
→ Archive @ 365d, dramatically reducing long-term archive cost.
- One auth pattern that works everywhere. Custom AzureSdkTokenProvider
handles Container Apps MSI, AKS Workload Identity, VM MSI, service
principals, and developer credentials with one config.
- Built-in observability. Azure Monitor Workbook with prebuilt KQL
queries for job activity, data volume, and cost.
- Productized FastAPI control plane with web UI for triggering jobs and
downloading manifests.

DEPLOYMENT

A single Bicep template provisions ADLS Gen2 + Key Vault + Log Analytics
+ Container Apps Environment (VNet-injected) + extract + restore Jobs +
Control Plane App. Marketplace UI walks you through the inputs; the
deploy takes ~3 minutes.

WHO IT'S FOR

- Regulated industries archiving Oracle for compliance retention while
needing occasional query access (HIPAA, PCI, government).
- Database teams sunsetting Oracle workloads who want a one-time
extraction that lives on as a queryable archive.
- Data platform teams seeding Azure data lakes from Oracle without
rebuilding the schema in Synapse or Fabric.

NOT for: real-time CDC replication (use Debezium/Oracle GoldenGate),
multi-master replication, or backup/recovery as a primary strategy
(use Oracle RMAN via Azure Backup for those).