The underlying server hardware is the single most important factor in determining the performance, scalability and availability of the VMs and the applications that run on them.
When deploying an enterprise database such as Microsoft SQL Server, one of the most critical tasks is deciding what hardware to deploy the system on. This is true even if you plan to deploy the database in a virtual machine (VM). Indeed, the underlying server hardware is the single most important factor in determining the performance, scalability and availability of the VMs and the applications that run on them.
HPE designed the Superdome X to provide the scalability and reliability required by the most demanding enterprise database applications. Here’s a look at some of the features in the HPE Superdome X that make it an exceptional server platform for mission-critical database implementations.
The Superdome X platform was designed from the ground up to provide extremely high levels of performance and scalability. It supports up to 384 cores and 24 TB of RAM. Proof of the system’s extreme scalability potential is demonstrated in several world-record-breaking benchmarks. Some of the benchmarks set by the HPE Superdome X include:
- No. 1 overall performance @ 10000 GB scale factor on non-clustered TPC-H benchmark1
- No. 1 overall eight-socket price/performance @ 10000 GB scale factor on non-clustered TPC-H benchmark2
- No. 1 x86 16P SPECjbb2015-MultiJVM max-JOPS and critical-JOPS benchmark results2
- x86 16P world record on the SPECjbb2015 Multi-JVM benchmark3
Reliability and availability are even more important than performance for a true enterprise database. Mission-critical databases require 24-by-7 availability and the level of advanced reliability, availability and serviceability (RAS) found in the HPE Superdome X. Taking full advantage of the Intel Xeon E7 v4 RAS capabilities, the HPE Superdome X’s RAS support includes:
- Processor RAS -- Superdome X servers use Intel’s Xeon E7 v4 processors, which include extensive capabilities for detecting, correcting, and reporting hard and soft errors. These RAS capabilities require the firmware support provided by the Superdome X; they are not supported in most other industry-standard servers.
- Memory RAS -- Main memory failures are a common causes of hardware downtime. The HPE Superdome X uses proactive memory scrubbing and Double Device Data Correction (DDDC) to improve memory availability. In addition, HPE Smart Memory DIMMs are designed to provide both performance and reliability.
- Platform RAS -- The HPE Superdome X offers built-in platform RAS features including Clock Redundancy, System Fabric RAS and Fault-Tolerant RAS.
- System Fabric RAS -- HPE Integrity Superdome X provides the fault-tolerant HPE Crossbar Fabric for connecting server blades.
For an overview of the fault management built into the HPE Superdome X, see Figure 1.
Figure 1 – HPE Superdome X fault management strategy
The HPE Superdome X implements a four-stage RAS strategy of detection, logging, analyzing and repair. This design minimizes the time to repair by capturing the correct data for diagnosis while still enabling the system to run to enable full error logging. The Superdome X provides component-level granularity for self-healing and field-replaceable components.
In addition to its RAS features, the HPE Superdome X provides hot-swappable power and cooling components. The Onboard Administrator’s (OA) built-in Analysis Engine also provides ongoing error-correction, self-healing and advanced diagnostics.
Finally, the HPE Superdome X’s hardware partitioning (nPar) enables organizations to implement multiple hardware partitions within the HPE Superdome X solution—essentially, enabling the system to function as if it were multiple separate server systems. These partitions are electrically isolated from one another, allowing users to separate different workloads in the different partitions. For example, an OLTP database application could be run in one partition and OLAP BI applications in another partition, with one having no impact on the other.
Choosing the proper server platform is the most essential element to ensuring successful implementation of mission-critical enterprise-level databases. The proper server platform ensures that organizations have the maximum levels of performance and availability, whether the database is running on bare metal or in a VM.
HPE and Microsoft are the underwriters of this article.