Composable vs Rack for a SAP HANA and VMware Refresh
A discrete manufacturer has two racks with 18 kW total power and no ability to add floor space. It runs 120 VMware VMs on 12 aging ProLiant DL380 servers and a 1.5 TB SAP HANA scale-up appliance. The SAP team requires HANA-certified infrastructure and will not virtualize HANA. The VMware team needs 35% growth over three years. The CIO has seen HPE Synergy and asks why the company should not use composable infrastructure for everything. A peer warns that Synergy is 'overkill and a single point of failure' because of the Composer. The refresh budget is capped at 8% above the current quote. The account team wants a single architecture to simplify support. A new 2.0 TB HANA sandbox is due in four months. The current HANA appliance is out of support in four months.
1. Recommend a platform mix and sizing guardrails. Explain where Synergy fits and where it does not.
Read the worked answer
Recommend a mixed platform, not one architecture for everything:
- SAP HANA: two certified ProLiant DL380 or DL560 nodes in a scale-up design, sized for 1.5 TB production plus a 2.0 TB sandbox. HANA certification and SAP support matter more than composability. Do not put HANA in a Synergy compute module unless the exact HANA certification and support matrix allow it.
- VMware: HPE Synergy with compute modules and image streams if density and profile mobility justify it; otherwise ProLiant DL380s with HPE OneView.
- Power/space: 18 kW and two racks are the binding constraints. Model kW per node and leave 20% headroom; do not chase 35% VM growth without a power review.
- Budget: if Synergy exceeds the 8% cap, use ProLiant for HANA and a smaller Synergy frame for VMware, or all ProLiant. Single support can be achieved through HPE Pointnext even with two platforms. The judgement: standardize operations, not necessarily hardware. Forcing HANA into Synergy to satisfy a slogan adds certification risk and may waste density.
2. How do you answer the peer's claim that Synergy is a single point of failure because of the Composer?
Read the worked answer
The Composer is a management plane, not the data path for running workloads. In a Synergy design, management modules are typically deployed as an HA pair, and existing compute modules continue to run if Composer is unavailable; provisioning and some profile changes pause. Ask the peer to compare that to the risk of 12 aging DL380s and a HANA appliance out of support. Mitigate with:
- HA Composer pair and documented backup/restore.
- A break-glass runbook for manual changes.
- Spare modules and support entitlements.
- A pilot that includes a Composer failure test. If the customer cannot accept any management-plane dependency, use ProLiant with OneView. The judgement is not whether Synergy is perfect; it is whether the operational risk is lower than the current estate and within the customer's tolerance.