Partner agencies and consulting firms across NZ and ANZ subcontract Oracle OCM and AWS PostgreSQL SME work to Dipak Rijal — NZ company, GST registered, NDA-ready, cleared to start fast.
Most agencies struggle to place database work that requires genuine depth. Here's what makes this partnership different.
Flexible arrangements to suit your project type, client relationship, and delivery preference.
| Service | Typical scenario | Demand | Duration |
|---|---|---|---|
| Oracle to PostgreSQL Migration | Enterprise moving off Oracle licensing to Aurora PostgreSQL on AWS | Very high | 3–12 months |
| AWS RDS / Aurora Architecture | Greenfield AWS database design, HA/DR, multi-region setup | High | 2–8 weeks |
| PostgreSQL Performance Rescue | Production queries running too slow, bloat, replication lag | High | 1–4 weeks |
| Oracle RAC / Data Guard / Exadata | Oracle enterprise HA/DR, upgrades, Exadata operations | Specialist | Variable |
| SQL Server Performance & HA/DR | Microsoft SQL Server tuning, Always On availability groups, migration to Azure SQL / AWS RDS | High | 1–8 weeks |
| AWS DMS Migration Delivery | Database migration using AWS DMS with full-load and CDC | High | 4–16 weeks |
| Database HA/DR Design | Disaster recovery architecture, RTO/RPO planning and testing | High | 2–6 weeks |
| AI-SDLC Delivery Advisory | AI-assisted delivery workflows for cloud database teams | Growing | Ongoing |
| Pre-sales / Tender Support | Technical input for database bids, SOW review, architecture decks | Retainer | Ongoing |
First conversation is free. Send the client context, database problem, delivery model or tender need. I will respond with a clear answer on fit, timeline and next step.
Direct enquiry delivered to partners@rijal.co.nz
Oracle OCM + AWS PostgreSQL SME + 15 years enterprise delivery. Available now. NZ company, GST registered.
Storage only — excludes instance hours, backups, data transfer. Approximate on-demand rates, verify on AWS calculator.
Rule of thumb only. Real-world adds LOB scans, type-conversion overhead, target write throughput, and validation passes.
Based on the classic Brandur/Bonesco formula: connections ≈ cores × multiplier. Cap at 200 unless you've measured otherwise.