PoC — Timeline
Choose the timeline that matches your PoC scenario. All three options use the same entry and exit criteria — they differ in scope, not rigour.
| Scenario | Timeline | Section |
|---|---|---|
| Scenario 1 — Single Tool Quick PoC | ~1 week | 1-Week Timeline |
| Scenario 2 — Two-Tool Comparison | ~3 weeks | 3-Week Timeline |
| Scenario 3 — Full Three-Tool Evaluation | 4–5 weeks | Full Evaluation Timeline |
1-Week Single-Tool PoC
Use when you have already chosen a tool and just need to validate the end-to-end process before production.
| Day | Activity |
|---|---|
| Day 1 | Deploy tool environment (Veeam, HYCU, or Carbonite management server); verify connectivity to Nutanix source; document 3–5 PoC VMs |
| Day 1–2 | Start initial backup, replication, or mirror job for all PoC VMs; record start time |
| Day 2 | Monitor initial copy completion; record duration; confirm incremental/delta cycle is healthy |
| Day 3 | (Two-hop paths only) Deploy Azure Migrate source appliance; register with project; discover Hyper-V staging VMs |
| Day 3 | Trigger final incremental; declare maintenance window; begin cutover — power off source VMs, run restore/failover/cutover |
| Day 3–4 | Validate all VMs using Hop 1 go/no-go checklist (see tool-specific validation.md); obtain batch sign-off |
| Day 4 | (Two-hop paths only) Start Azure Migrate replication; wait for Protected state; run test migration on isolated network |
| Day 4–5 | Azure Migrate production cutover; validate on Azure Local; record Hop 2 downtime |
| Day 5 | Rollback drill: roll one VM back to Nutanix; record duration; complete final Go/No-Go sign-off |
Minimum deliverables: Timing log, per-VM gate checklist, one rollback drill record, signed Go/No-Go.
3-Week Two-Tool Comparison
Use when comparing two tools side-by-side with the same workload set.
Week 1 — Setup and Tool 1
| Day | Activity |
|---|---|
| Day 1 | Deploy both tool environments; verify connectivity to Nutanix source; document 5 shared PoC VMs |
| Day 1–2 | Tool 1: Start initial backup/replication/mirror for all 5 PoC VMs; record start time |
| Day 2 | Monitor Tool 1 initial copy; record duration; confirm incremental cycle |
| Day 3 | Tool 1 Cutover: maintenance window; final incremental; power off source VMs; restore/cutover to staging (Hyper-V or Azure Local) |
| Day 3–4 | Validate Tool 1 VMs using Hop 1 checklist; record per-VM downtime |
| Day 4–5 | (Two-hop paths) Azure Migrate replication for Tool 1; test migration; record Protected state timing |
| Day 5 | Tool 1 Azure Migrate production cutover; validate on Azure Local; record Hop 2 downtime |
Week 2 — Tool 2
| Day | Activity |
|---|---|
| Day 6 | Power source VMs back on (restored to last-known state on Nutanix); re-baseline before Tool 2 run |
| Day 6–7 | Tool 2: Start initial backup/replication/mirror for the same 5 PoC VMs; record start time |
| Day 7–8 | Monitor Tool 2 initial copy; confirm incremental cycle |
| Day 9 | Tool 2 Cutover: maintenance window; final incremental; power off source VMs; restore/cutover |
| Day 9–10 | Validate Tool 2 VMs using Hop 1 checklist; record per-VM downtime |
| Day 10 | (Two-hop paths) Azure Migrate replication for Tool 2 VMs |
Week 3 — Comparison and Decision
| Day | Activity |
|---|---|
| Day 11–12 | Tool 2 Azure Migrate test migration and production cutover; validate on Azure Local; record Hop 2 downtime |
| Day 13 | Rollback drill for each tool (one VM each); record duration |
| Day 14 | Complete comparison scorecard (see Scenario 2 Scorecard); review weighted results |
| Day 15 | Decision meeting — select production tool; draft production wave plan; sign Go/No-Go |
Minimum deliverables: Timing log for both tools, comparison scorecard, one rollback drill per tool, signed recommendation and Go/No-Go.
Full Evaluation Timeline
Original 4–5-week plan for Scenario 3 (all three tools, both staging options).
Week-by-Week Plan
Week 1 — Environment Setup
| Day | Activity |
|---|---|
| Day 1–2 | Deploy HYCU controller VM on Nutanix AHV cluster |
| Day 1–2 | Deploy Veeam Backup & Replication server |
| Day 1–3 | Deploy Carbonite management server; validate connectivity to Nutanix source and Azure Local target; confirm licensing |
| Day 2–3 | Configure HYCU source, backup target, and Hyper-V restore target |
| Day 2–3 | Configure Veeam source, AHV proxy, and Hyper-V replication destination |
| Day 3–4 | Provision standalone Hyper-V staging host for Option A tests |
| Day 3–5 | Verify Azure Local cluster health, CSV capacity, and Azure integration readiness |
| Day 4–5 | Deploy Azure Migrate appliance; register with project |
| Day 5 | Identify and document the 15 PoC VMs; record baseline OS, apps, IPs, and services |
Week 2 — Option A: Cells A1 + A2
| Day | Activity |
|---|---|
| Day 6–7 | Cell A1: Start Veeam replication job (PoC-VM-01..05) to standalone Hyper-V |
| Day 6–7 | Cell A2: Start HYCU full backup (PoC-VM-06..10) |
| Day 7 | Monitor initial full copy completion; record timing |
| Day 8 | Let incremental cycles run; confirm CBT or incremental behavior |
| Day 9 | Cell A1 Cutover: final sync, failover, re-IP validation |
| Day 9 | Cell A2 Cutover: final increment, restore, post-restore re-IP validation |
| Day 10 | Validate A1 and A2 VMs on Hyper-V and start Azure Migrate replication |
Week 3 — Option A: Cell A3 + Azure Migrate Completion
| Day | Activity |
|---|---|
| Day 11–12 | Cell A3: Deploy Carbonite agents on PoC-VM-11..15 source and pre-provisioned Azure Local targets; start initial mirror |
| Day 12–13 | Monitor initial mirror completion; confirm continuous replication; validate delta lag < 60 seconds |
| Day 13 | Test migration for A1/A2 workloads to isolated Azure Local network; validate |
| Day 14 | Production cutover for A1/A2 workloads; collect evidence and metrics |
| Day 14–15 | Cell A3 Cutover: Carbonite production cutover for PoC-VM-11..15; validate on Azure Local; record downtime |
| Day 15 | Rollback drill for one representative workload; prepare Azure Local-hosted Hyper-V staging for Option B |
Week 4 — Option B: Cells B1 + B2
| Day | Activity |
|---|---|
| Day 16–17 | Cell B1: Re-run Veeam workload set using Azure Local-hosted Hyper-V staging |
| Day 16–17 | Cell B2: Re-run HYCU workload set using Azure Local-hosted Hyper-V staging |
| Day 18 | Validate both cells on staging and monitor Azure Local resource impact |
| Day 19 | Run Azure Migrate test migration for B1 and B2 |
| Day 20 | Complete B1 and B2 cutovers; document capacity, throughput, and operational complexity |
Week 5 — Option B: Cell B3 + Decision
| Day | Activity |
|---|---|
| Day 21–22 | Cell B3: Re-run Carbonite workload set or alternate workload archetypes against Azure Local targets |
| Day 23 | Validate B3 workloads; record Azure Local resource impact under concurrent Carbonite + Azure Migrate load |
| Day 24 | Complete scorecard, finalize risk register, and compare all cells |
| Day 25 | Go/No-Go meeting — select the production migration tool and staging model |
PoC VM IP / Naming Assignments
Use IIC naming convention for PoC VMs on Azure Local:
| PoC VM | Source Name | AZL Target Name | Source IP | Target IP |
|---|---|---|---|---|
| PoC-VM-01 | <nutanix-vm-name> | <azl-vm-name> | <ip> | <ip> |
| PoC-VM-02 | — | — | — | — |
| ... | — | — | — | — |
Fill in this table before starting Week 2. Shared with networking team for VLAN/subnet changes.
Rollback Runbook (PoC)
Rollback triggers
- Tier 1 VM fails application smoke test after cutover
- Data integrity mismatch detected (checksums/record counts/files)
- Cutover exceeds agreed downtime window
Rollback steps
- Declare rollback and freeze further cutovers in current cell
- Stop target workload on Azure Local VM
- Power on source VM on Nutanix (or restore prior running state)
- Repoint DNS/LB records to source VM
- Validate app health from business-owner test script
- Record rollback duration and root cause in metrics workbook
Rollback validation
- Source VM serving production traffic
- Application owner confirms service restoration
- Incident notes and remediation plan documented before next test wave
Capacity and Saturation Thresholds
Treat these as stop-wave thresholds during PoC execution:
| Domain | Threshold | Action if exceeded |
|---|---|---|
| Azure Local node CPU | > 85% sustained for 15+ minutes | Pause replication/cutover wave; scale down concurrency |
| Azure Local memory pressure | < 15% free sustained | Pause non-critical test jobs |
| CSV free capacity | < 25% remaining | Stop new replication; reclaim space |
| Storage latency | > 20 ms sustained | Pause wave; investigate storage bottleneck |
| Replication network throughput | < planned floor for 30+ minutes | Extend window and re-baseline estimate |
Decision Framework
After the PoC, answer these questions to finalize tool selection:
| Question | Answer | Impact |
|---|---|---|
| Did all three tools successfully migrate their workload sets? | Yes / No / Partial | Eliminate failing tools from production shortlist |
| Which tool had the fastest initial transfer or restore window? | Veeam / HYCU / Carbonite | Performance matters for wave planning |
| Which tool had faster cutover window? | < 30 min per VM preferred | |
| Did re-IP work reliably end-to-end? | Veeam has advantage if complex re-IP needed | |
| Did HYCU backup target add complexity? | If yes, consider Veeam or Carbonite | |
| Did Carbonite replication lag stay consistently low before cutover? | If not, investigate network or storage bottleneck | |
| Was standalone HV needed, or is Azure Local-hosted Hyper-V sufficient? | If Azure Local is sufficient, reduces hardware | |
| Team preference after hands-on use? | Operator comfort matters at 300-VM scale |
Weighted scorecard model
Use weighted scoring to prevent subjective decisions:
| Category | Weight | A1 | A2 | A3 | B1 | B2 | B3 |
|---|---|---|---|---|---|---|---|
| Migration fidelity / reliability | 30% | ||||||
| Speed (full + incremental + cutover) | 20% | ||||||
| Operational complexity | 15% | ||||||
| Re-IP and network stability | 10% | ||||||
| Rollback performance | 10% | ||||||
| Capacity efficiency / hardware footprint | 10% | ||||||
| Team fit / supportability | 5% |
Scoring method:
- Rate each category 1-5 per cell
- Weighted score =
rating * weight - Sum weighted scores to rank A1-A3 and B1-B3
- Any auto-fail gate overrides score and disqualifies that cell
Outcome Scenarios
| Outcome | Production Recommendation |
|---|---|
| HYCU is fastest and simplest for AHV workloads | Use HYCU with the best passing A/B cell |
| Veeam has best cutover and re-IP behaviour | Use Veeam with the best passing A/B cell |
| Carbonite has lowest downtime and no staging hardware needed | Use Carbonite (Deploy-First) path |
| Azure Local-hosted Hyper-V adds no instability | Use the best passing B-cell — reduces hardware |
| Azure Local staging causes capacity issues | Use the best passing A-cell |