Cloud Infrastructure · Leadership Update

Moving PAD to Azure Cloud retires legacy infrastructure and unlocks faster, more reliable, and lower-cost releases.

Database replication to the Azure testing environment is complete as of June 10. End-to-end testing begins in July — with a full production release targeted for September, or sooner if testing resolves cleanly.

What this migration unlocks
Hardware costs scale with actual usage — not fixed purchase decisions and recurring refresh cycles
Reduced on-prem maintenance and support spend
Faster time to market on application updates
More frequent, more flexible release cycles
Fewer critical failures and P1 incidents
Current Status & Milestones

Both layers are live in Azure. Testing in July — production target September or sooner.

The migration is complete in the testing environment. Attention now shifts to comprehensive end-to-end validation before promoting to production.

June 10, 2025 DB Replication Complete
July 2025 End-to-End Testing
Sept 2025 or sooner Production Release
Complete

Application Layer

PAD Admin and all supporting APIs are running in Azure. The cloud application tier is fully operational and consuming Azure-hosted data.

✓ Complete
Complete

Database Replication

PAD databases successfully replicated from on-prem SQL Server environments to Azure SQL Managed Instance across DEV, TEST, and STAGE.

✓ Complete — June 10, 2025
Active — July 2025

End-to-End Testing

Comprehensive validation across application and database layers — ensuring full system integrity before the production release. Any adjustments identified during testing are resolved before PROD promotion.

  • Application layer: PAD Admin, Document Services, and all APIs against Azure-hosted data
  • Database integrity across DEV, TEST, and STAGE environments
  • End-to-end transaction and workflow verification under realistic load
  • Full regression suite — no regressions accepted before PROD promotion
Executive message: The application and database layers are both live in Azure as of June 10. End-to-end testing begins in July to validate full system integrity — with a production release targeted for September, or earlier if testing resolves cleanly.
Test Strategy

Comprehensive end-to-end testing underway — July 2025

Two parallel test tracks — application layer and database layer — run concurrently to validate full system integrity before the production release.

Application Layer Testing

1

PAD Admin Console Validation

Full functional verification of the Azure-hosted admin application — all screens, workflows, and user roles against live Azure-connected data.

2

API Integration Testing

Verify all supporting APIs return correct responses with Azure SQL MI as the data source. Validate request/response contracts, error handling, and timeout behavior.

3

Document Services End-to-End

Confirm document generation, retrieval, and operational workflows function correctly against Azure-hosted data across all environments.

4

Regression Suite

Full automated regression covering all critical PAD user journeys. No regressions accepted before PROD promotion.

Database Layer Testing

1

Data Integrity Verification

Row-count validation and schema comparison between on-prem source and Azure SQL MI across DEV, TEST, and STAGE. Zero data loss tolerance.

2

Transaction Log Chain Validation

Verify continuity of the transaction log chain post-migration. Confirm no gaps or corruption exist before promoting to PROD.

3

Query Performance Benchmarking

Run representative query workloads against Azure SQL MI and compare execution plans and latency against on-prem baseline. Address any regressions before release.

4

Failover and Recovery Testing

Validate Azure SQL MI high-availability and failover behavior. Confirm RTO/RPO targets are met and backup/restore procedures are documented and tested.

Migration Architecture

Application and database — both now in Azure

The full migration path is complete through the testing environment. All four stages are operational; PROD promotion follows successful end-to-end testing.

1

On-Prem PAD SQL Environments

DEV6 HOTFIX OPSFIX QATEST TEST STAGE PROD
Source databases on on-prem SQL Servers — replicated to Azure SQL MI as of June 10.
2

Connectivity & Migration Path

VPN / network connectivity
Azure firewall / secure routing
SHIR VM
Azure Database Migration Service (DMS)
Backup share access (.bak / .trn)
All runbook steps complete — replication path fully operational as of June 10.
3

Azure Cloud Application Layer

PAD Admin Console Running in Azure — APU / AMU administrative site, fully operational
Document Services Operational workflows available in Azure, connected to Azure-hosted data
Hosted in APUS / APEI Azure environment — Management Group: apei-global-sandbox
4

Target State

Shared Azure SQL Managed Instance
PROD database migrated after E2E sign-off
PAD APIs fully on Azure-hosted data
PAD Admin fully cloud-hosted end to end
On-prem SQL servers repurposed or decommissioned

Where we stand: Application and database replication are both complete in the testing environment. E2E testing in July clears the path to full production parity — on-prem dependency retired.

On the Horizon

Broader cloud strategy: the right partner for the long term

PAD migration is the near-term priority. In parallel, a discovery and assessment phase is underway to determine APEI's long-term cloud direction.

The current Azure footprint is overly complex and piecemeal, with significant automation gaps and no foundational strategy. Before further build-out, APEI is evaluating cloud partners against cost, automation, and long-term fit.

  • Engage cloud partners to identify an optimal long-term strategy
  • Quantify cost savings and total cost of ownership
  • Enable rapid feature deployment through modern, automated architecture
  • Migrate all workloads to full CI/CD automation from day one
Microsoft Azure Current State — In Review

APEI's current cloud environment. The PAD migration is completing here, but the broader Azure footprint lacks architectural coherence. Under review before any further workload investment.

Current limitations driving the review
  • Overly complex, piecemeal architecture with no foundational strategy — grown organically rather than designed
  • Significant automation gaps across the environment; CI/CD coverage is incomplete
  • Microsoft Copilot and AI integrations lag behind AWS and Google in maturity and native capability
  • Highest per-unit cost among the three platforms at comparable workload sizes
AWS Strong Candidate

The market leader in cloud infrastructure. AWS offers the broadest service catalog, the most mature automation tooling, and consistently competitive unit economics — making it the strongest candidate for a long-term platform.

Advantage vs. Azure
  • Material cost savings through right-sized, on-demand architecture — hardware costs scale with actual usage, not committed spend
  • Deepest CI/CD and DevOps toolchain in the market — full automation from day one is standard, not aspirational
  • More mature autoscaling and workload optimization capabilities reduce infrastructure overhead
  • Rapid agentic lift-and-shift with cleaner architectural patterns from the outset
  • Largest global partner ecosystem — broadest access to expertise, tooling, and integrations
Google Cloud High Potential

Google Cloud brings competitive pricing, strong data and analytics capabilities, and the most native AI integration of any platform — particularly compelling given APEI's growing AI strategy.

Advantage vs. Azure
  • Competitive pricing with lift-and-shift capability — strong unit economics especially at data and analytics scale
  • Gemini AI is natively integrated across the full Google Workspace suite, enabling AI-assisted workflows without additional licensing or configuration
  • BigQuery and integrated data layer make it the strongest fit for analytics-heavy workloads and data platform consolidation
  • Migrating to Google Workspace could replace Microsoft 365 — AI-native productivity tools across email, docs, and collaboration from day one
  • Strong Kubernetes-native infrastructure (GKE) reduces container overhead and supports modern, portable application architecture