Data Sovereignty and Burst Capacity in a GreenLake Deal
A regional bank runs 400 VMware VMs and 80 TB usable on HPE ProLiant and Nimble in its primary data center, with a DR site 200 km away. Regulators require data residency and a tested DR plan. The bank wants to use public cloud for analytics burst but not for core banking. Growth is uncertain: 5% baseline, possible 25% if a merger closes. The CIO has a three-year TCO target and will consider HPE GreenLake, but the CFO fears overcommitting to a minimum. HPE Pointnext has been asked to quote deployment and a DR test. The incumbent competitor claims its consumption model is 12% cheaper. The merger decision is nine months away. The CIO will not sign a three-year commitment without a downside case.
1. How would you structure the GreenLake commitment and buffer so the CFO is protected if the merger does not close?
Read the worked answer
Structure the deal around measured baseline, not merger speculation:
- Baseline: 400 VMs and 80 TB usable with 5% growth. Commit to the measured steady-state plus a 15-20% buffer, not the 25% merger case.
- Buffer: use GreenLake burst capacity or a pre-agreed true-up rather than overcommitting for nine months. Set a 90-day review and a documented growth trigger.
- Merger: keep a separate optional expansion schedule; if the merger closes, add capacity under the same terms. Do not lock the CFO into the full 25% today.
- Public cloud: use for analytics burst only, with data movement and egress limits. Keep core banking in the data center for residency and latency.
- Exit: require data export, standard formats, and termination assistance. The judgement: a lower commit with a clear expansion path is better than a large discount that creates unused capacity. Price the buffer explicitly.
2. The competitor claims 12% lower cost. How do you build and present a defensible TCO comparison?
Read the worked answer
Normalize the competitor's 12% claim before responding. Build a three-year TCO with the same scope:
- Compute/storage capacity, support, and software licensing.
- Power, floor space, and refresh avoidance.
- Migration, dual-running, training, and Pointnext deployment.
- DR site, replication, and tested DR runbook.
- Public cloud egress and analytics burst.
- Risk: unused commit, penalty for overage, exit cost. If the competitor is truly 12% cheaper on a like-for-like basis, do not match blindly; quantify what is missing (support tier, DR test, migration). Position GreenLake on consumption flexibility and measured billing. Use a three-year NPV and a sensitivity case for 5% vs 25% growth. The CFO's real fear is overcommitment, so show the downside case, not just the best case.
3. Define the Pointnext services scope and the acceptance criteria for the DR test.
Read the worked answer
Define Pointnext scope as outcomes with acceptance criteria:
- Deployment: install, configure, integrate with existing network and backup, and document.
- Migration: wave plan for 400 VMs, rollback per wave, and application owner sign-off.
- DR: build runbook, execute a full test from the 200 km site, measure RTO/RPO, and fix gaps.
- Knowledge transfer: runbooks, training, and 30-day hypercare.
- Out of scope: application remediation, data cleanup, and merger integration. Success criteria: DR test meets the regulatory RTO, no P1 during migration, and the bank's team can operate the runbook. Tie payment to those gates. This protects both sides and prevents services from becoming open-ended.