Home > Workload Solutions > SAP > Guides > SAP HANA TDI Guides > Dell Validated Design for SAP HANA TDI with Dell PowerMax 2000 and 8000 Storage Arrays > Dell EMC PowerMax storage
The PowerMax family is the first Dell EMC data storage product to fully use NVMe technology for customer application data. The PowerMax array is built using an end-to-end NVMe storage architecture, allowing it to reach unprecedented I/O densities and performance by eliminating the flash media choke points found with traditional SAS and SATA interfaces. With the PowerMax array, customers can deploy innovative applications in the areas of real-time analytics, machine learning, and big data, which demand lower latency and higher performance than legacy all-flash offerings provided.
The Dell EMC PowerMax family consists of two models—the PowerMax 2000 and the flagship PowerMax 8000—as shown in the following figure. The PowerMax 2000 is designed to provide Dell EMC customers with efficiency and maximum flexibility in a 20U footprint. The PowerMax 8000 is designed for massive scale, performance, and IOPS density, all within a one- or two-floor-tile footprint.
Figure 2. Dell EMC PowerMax product family
Both PowerMax models have at their foundation the Dynamic Virtual Matrix architecture and PowerMaxOS, which was created for the NVMe platform. PowerMax systems are true all-flash arrays—that is, they are targeted to meet the storage capacity and performance requirements of the all-flash enterprise data center. The feature-rich PowerMax products are designed to take advantage of ultrahigh-performing storage class memory (SCM) and high-capacity standard flash drives to create the densest storage configuration possible.
The PowerMax platforms offer customers these main benefits:
Note: For more information about the Dell EMC PowerMax family, see the Dell EMC PowerMax: Family Overview White Paper.
The following table shows the system specifications for the PowerMax 2000 and 8000 arrays:
Table 1. PowerMax system specifications
PowerMax model | Bricks per model | Brick CPU | Brick cache | Maximum cache | Maximum number of front-end modules per Brick | Maximum number of drives | Maximum capacity |
2000 | 1-2 | 12 core, 2.5 GHz Intel Broadwell (2 per director)
Total: 48 cores per Brick | 512 GB, 1 TB, and 2 TB (DDR4) | 4 TB | 8 (32 total front-end ports per Brick)
Total: 64 front-end ports per system | 96 | 1 PB |
8000 | 1-8 | 18 core, 2.8 GHz Intel Broadwell (2 per director)
Total: 72 cores per Brick | 1 TB and 2 TB (DDR4) | 16 TB | 6 (24 total front-end ports per Brick—open systems/mixed)
Total: 256 front-end ports per system | 288 | 4 PB |
PowerMax configurations consist of modular building blocks called PowerMax Bricks. The modular Brick architecture reduces complexity and allows for easier system configuration and deployment. This architecture also allows the system to scale while continuing to deliver predictable high performance.
Two types of Bricks are available for PowerMax configurations:
The Brick concept allows PowerMax platforms to scale up and scale out. Customers can scale up by incrementally adding Flash Capacity Packs. For the PowerMax 8000, each Flash Capacity Pack has 13 TB of usable storage. For the PowerMax 2000, each Flash Capacity Pack has either 11 TB or 13 TB of usable storage, depending on the RAID protection type selected. PowerMax systems scale out by aggregating up to two Bricks for the PowerMax 2000 and up to eight Bricks for the PowerMax 8000. Scaling out a PowerMax system by adding Bricks produces a predictable, linear performance improvement regardless of the workload.
Note: For detailed information about available PowerMax Brick configurations, see the PowerMax Family Specification Sheet.
PowerMax 2000 and 8000 systems support storage class memory (SCM) drives. These drives are powered by dual-port Intel Optane technology, together with machine learning algorithms, for data placement when both NVMe and SCM drives are used. SCM drives are high-performance drives that have a significantly lower latency than NVMe flash drives, thus increasing response times.
A PowerMax model that contains both types of drives effectively has two storage tiers—the higher-performance SCM drives and the NVMe drives. Automated data placement takes advantage of the performance difference to optimize access to data that is frequently accessed. Automated data placement can also help to optimize access to storage groups that are used for workloads such as SAP HANA that have higher-priority service levels.
For optimum cost per performance, we recommend that the total usable capacity (TBu) of SCM Tier 0 is between 3 and 12 percent of the required effective capacity (TBe) of the system.
PowerMax arrays use the PowerMax Adaptive Compression Engine (ACE) for inline hardware compression. The ACE data-reduction method provides the highest space-saving capability with a negligible performance impact. PowerMax technology uses inline hardware-based data deduplication, which identifies repeated data patterns on the array and stores the repeated data only once. This functionality prevents the consumption of critical core resources on the PowerMax system and limits performance impact.
Note: For information about PowerMax data services and how to configure and use SCM drives on PowerMax arrays, see the Dell EMC PowerMax: Family Overview White Paper.