Hardware
Hardware sizing determines whether an SAP system performs and scales, or struggles. SAP provides tools and rules to translate business requirements into CPU, memory, disk and network needs, and HANA raises the bar significantly.
In the cloud, sizing maps to certified instance types.
- SAP sizing uses the SAP Quick Sizer, which converts business inputs (number of users, transaction volumes, data size) into hardware…
- HANA is memory-resident and must run on certified hardware or certified cloud instances, with strict CPU-to-memory ratios.
- CPU (SAPS): for the application and database tiers.
- Watch out: Under-sizing HANA memory, the dataset must fit.
How SAP sizing works
SAP sizing uses the SAP Quick Sizer, which converts business inputs (number of users, transaction volumes, data size) into hardware requirements expressed in SAPS (SAP Application Performance Standard) for CPU, plus memory and disk. This gives a vendor-neutral basis for procuring or provisioning hardware.
The dimensions to size
- CPU (SAPS): for the application and database tiers.
- Memory: critical for HANA, which holds data in RAM.
- Disk/storage: for the database, logs, and backups.
- Network: low latency between app and database tiers.
HANA’s special requirements
HANA is memory-resident and must run on certified hardware or certified cloud instances, with strict CPU-to-memory ratios. Memory sizing is especially important: the whole (or hot) dataset must fit in RAM. This is a major difference from traditional disk-based databases.
Cloud sizing
In the cloud, sizing maps to certified instance types. The same SAPS/memory logic applies, but you choose from certified VM shapes rather than physical servers.
Common pitfalls
- Under-sizing HANA memory, the dataset must fit.
- Uncertified hardware/instances for HANA.
- Ignoring network latency between tiers.