Home > Workload Solutions > SAP > Archive > Enterprise Content Management Solution for S/4HANA Migrations > Data tiering
When considering your new infrastructure requirements, think of a data tiering strategy that is based on the concept of data temperature: hot, warm, and cold. You can reduce your S/4 HANA licensing costs by using Dell Technologies storage systems for warm and cold data and not keeping all your data in memory.
In SAP terms, “hot data” is mission-critical data for real-time processing and analytics that is stored in server memory. SAP has certified four-socket Dell PowerEdge servers for tailored data center integration (TDI) deployments. On the R840, R940, and MX840c PowerEdge models, up to 6 TB of data can be stored in DRAM. If more than 6 TB are needed in an SAP HANA TDI deployment, you can extend the hot memory tier to 15 TB with Intel Optane DC Persistent Memory in App Direct mode.
When you add Intel Optane Persistent Memory DIMMs, SAP HANA controls what is placed in persistent memory and what remains in DRAM. Volatile data structures stay in DRAM while column-store data, which can be more than 90 percent of SAP HANA data in many systems, is moved to persistent memory. This approach is less costly than buying an eight-socket server and putting all your SAP HANA data in DRAM. It also increases the overall S/4 HANA data capacity on your PowerEdge servers.
SAP HANA SPS 04 introduced SAP HANA Native Storage Extension (NSE), a feature that enables you to specify certain data as warm or “page loadable.” This data is loaded page by page into memory as required for query processing. Unlike column-loadable data, page-loadable data does not need to reside completely in memory, so its memory footprint is smaller. The SAP HANA database is partly in memory and partly on disk, as shown in the following figure:
PowerMax storage arrays can greatly increase SAP HANA data capacity and reduce the overall TCO for customers. This Dell Technologies hardware platform is among the first to use an end-to-end Non-Volatile Memory Express (NVMe) architecture for customer data. A machine-learning engine is built into PowerMax arrays for automated data placement between cache, NVMe flash drives, and NVMe storage class memory (SCM) drives.
An ideal way to reduce the total cost of ownership (TCO) of your S/4 HANA infrastructure is to archive as much nonactive (or “cold”) data as possible. Typically, this data is historical data that must be kept for long periods of time for legal or regulatory reasons. Isilon scale-out NAS systems, specifically the A200 and A2000 models, are designed for archival storage. The A200 model (120-720 TB per chassis) provides nearline storage for active archival storage, while the A200 is ideal for high-density, deep archival storage. Cold data is managed separately from the SAP HANA database but is still accessible at any time through SAP HANA’s data federation capabilities.